Deciphering Metal-Impacts on Synthetic and Natural Microbial Communities

Proteomics and Microbiology Faculty of Sciences University of Mons, UMONS Mons, Belgium

Deciphering metal-impacts on synthetic and natural microbial communities

Dissertation submitted in fulfilment of the requirements for the degree of Doctor of Philosophy in Biological Sciences by Valentine Cyriaque

This work was promoted and funded by the Fond National de Recherche Scientifique (F.R.S.-FNRS)

Jury: Pr. Ruddy Wattiez –UMONS Director of Thesis Pr. David C. Gillan– UMONS Co- Director of Thesis Pr. Patrick Flammang – UMONS Academic President of the Jury Pr. Søren Sørensen – University of Copenhagen External Reviewer Pr. Dirk Springael – KU Leuven External Reviewer Pr. Isabelle George – ULB External Reviewer

Cover: Tyson Dudley

“Don't waste your hate Rather gather and create Be of service, be a sensible person Use your words and don't be nervous You can do this, you've got purpose Find your medicine and use it” Nahko, Manisfesto

“Sono pessimista con l'intelligenza, ma ottimista per la volontà.” “Pessimiste par l’intelligence mais optimiste par la volonté.” “Pessimistic by intelligence but optimistic by the will.” Antonio Gramsci, Prison letter (1929)

Summary Metal contamination poses biotoxicity and bioaccumulation issues, affecting both abiotic conditions and biological activity in ecosystems. The use of metals and metalloids as raw materials, in industries and technologies drastically increased from the industrial revolution and urbanization of the XVIIIth century. For 100 years (1893-2003), the MetalEurop foundry released zinc, copper, cadmium and lead directly into the river “la Deûle”, resulting in up to 30- fold increase in metal concentrations in downstream sediments. We used an integrative approach coupling in-situ 16Sr RNA sequencing from both DNA and RNA extracts, microcosm supervision and Horizontal Gene Transfer (HGT) monitoring in order to fully understand the mechanisms driving community resilience to metal pollution. We applied the ecological concept of Functional Response Groups (FRGs) to decipher the adaptive tolerance range of the in-situ sediment communities through characterization of microbial strategists, revealing differences in diversity and composition. Furthermore, in-vitro microcosm analysis with upstream non-polluted sediments challenged with metals and daily supplied of fresh river water, allowed to monitor the short-term impact of metal pollution on the microbial community over 6 months, in controlled conditions. We used qPCR and 16S rRNA gene amplicon sequencing with the ecological concept of Treatment Response Groups (TRGs). Both in-situ and microcosm studies reinforced the notion that mechanisms are taking place at the community level to face the metal pollution, such as facilitation processes and microbial community coalescence leading to an unexpectly high microbial diversity. However, if in-situ results suggest HGT as a key process in the long-term resilience of the community, the monitoring of IncP plasmid by qPCR in microcosms revealed a negative effect of metals on the short-term in comparison with control microcosms. Deeper HGT analysis using qPCR revealed the presence of an enriched native pool of conjugative IncP plasmids in the MetalEurop polluted sediments, confirming their importance for long-term adaptation of the community facing metal contamination. Furthermore, in-vitro conjugation assays coupled to Fluorescence Activated Cell Sorting (FACS) allowed to assess the plasmid transfer rate and permissiveness in Férin and MetalEurop communities showing sediments as hotspot for plasmid transfer. To understand the direct metal-impact on plasmid persistence in a microbial community, we used a consortium conjugative assay to monitor the dispersion efficiency of the plasmid in the recipient strain population by flow-cytometry and link it to the metaproteomic profile of the consortium by SWATH quantitative proteomics.

Acknowledgements

When time comes to write acknowledgements, the thesis can be considered with the benefit of hindsight. When I attended my first class in biology, I’d never thought I would now dive into the “micro”. Happy thing, I have now realized how bacteria are fascinating (and I hope this work will convince you about that). This PhD was the continuity of 5 years of permanent exploration and celebration of life in all its forms. Nine years in the scientific community, ups and downs were deeply rewarding, scientifically and humanely.

Then, first and foremost, I would express my deep gratitude to my supervisor, the Pr. Ruddy Wattiez, “Chef”, who gave me the opportunity of doing this PhD. Thank you so much for your support and your precious time during these 4 years.

I would thank my co-supervisor, Pr. David Gillan for his very careful readings.

Of course, I would thank the ProtMic team for these years spent together. To Melanie for having watched over my work during my master thesis and beyond, for having taught me the tricks! To Augustin, as a master student first, and as a colleague. To Giuseppe, for all these hours spent at the bench with me combined with joyful conversations. To Cyril, for his help at the bench and long anticipatory conversations. To Baptiste and Corentin for their help with mass spectrometry. To Catherine for making sure everything was possible! To Clotilde, Quentin, Fred, Jérémy and Samia who integrated me into the team. To Guillaume, Neha, Corentin and Kurt for the fun in the office. To Alice, Camille, Angela, Haixia, Paloma, Camille, Ghenima, Catherine, Baptiste, Vincent, Médéric, Grégoire, for the daily good mood, fun and kindness!

I would also express my thankfulness to Pr. Søren Sørensen for welcoming me in his lab. I was a fresh PhD student back then, and these 6 months have been rich in experience and encounters. I would also then thank the section of microbiology, Mette, Anders, Leise, Luma, Anette, Stephan, Tim, Trine, Ines, Whenzeng, Martin, Sten, Annelise, Whenzeng, Henriette, and Jakob for having warmly welcomed me in their team. Special thanks to Jonas for mentoring me during these 6 months, to Dingrong, Shaodong, Milena, Rafa, Jakob and Urvish, colleagues and friends, and to Sam for being a mentor and taking me under his wing.

I would also thank the LASIR team, especially Gabriel Billon for the help with sediments, and the GIGA, especially Laurence Fievez, for the help with the flow cytometry. Thank you both for your availability, professionalism and kindness

I would also express my gratitude to the members of the jury who carefully reviewed this work.

I wish to thank the FNRS fellowship for the financial support.

Pendant mon parcours universitaire, j’ai eu la chance de suivre des cours donnés par des profs passionnants. Parmi eux, je voudrais remercier le Pr. Rasmont et sa passion contagieuse pour l’écologie.

Je voudrais remercier Julie et Baptiste pour les fins de rapport nocturnes, les blocus et les échanges d’insectes. Ces 5 années n’auraient pas été pareilles sans l’entraide, l’amitié et les rires. Merci à Pierre, Mathilde, Edwicka et Geoffrey.

Comprendre la vie est probablement le plus beau métier du monde.

“First of all, the beauty that he [the artist] sees is available to other people and to me, too, I believe, although I might not be quite as refined aesthetically as he is. But I can appreciate the beauty of a flower.

At the same time, I see much more about the flower that he sees. I could imagine the cells in there, the complicated actions inside which also have a beauty. I mean, it's not just beauty at this dimension of one centimetre: there is also beauty at a smaller dimension, the inner structure... also the processes.

The fact that the colours in the flower are evolved in order to attract insects to pollinate it is interesting - it means that insects can see the colour.

It adds a question - does this aesthetic sense also exist in the lower forms that are... why is it aesthetic, all kinds of interesting questions which a science knowledge only adds to the excitement and mystery and the awe of a flower.” Richard Feynman, 1981.

Apprendre devient alors tellement excitant. A chaque étape, des noms restent. Merci à Mme Gaëtane, Mme Dewinck, Mme Blondel, Mme Vandromme et Mme Wanuffel.

A mes amis. Ils sont ces lanternes colorées sur mon chemin. Merci à Sarah, Julie, Prési, Nico, Eme, Thom, Galaad, Dilan, Do, Pierre, Simon, Théo, Camille, Anto, Matthieu, Baptiste. A Antchoiine (Viens on y va !).

III

Last but not least, Papa, Maman, merci pour votre soutien inestimable, votre amour et votre fierté qui me poussent en avant. Merci pour les balades en forêt, pour les livres, pour avoir cultivé ma curiosité ! A Alex pour la patience, pour les rires et les jeux.

Aloha

Table of Content

Summary…………………………………………………………………………….…………... I Acknowledgements……………………………………………………………………………… II Table of content……………………………………………………………………….………… IV Publication, oral communications and poster presentation……………………………………… VII Abbreviations…………………………………………………………………….………….…... IX General Introduction 1 1. Foreword……………………………………………………………………………………… 3 2. The river ecosystem and its sediment microbiome …………………………...... 7 3. Biofilms in river ecosystems………………………………………………………...... 9 4. Metal contaminated environments and how bacteria deal with them………………………… 9 4.1. Metals, definition and classification…………………………………...... 11 4.2. Metals in aquatic ecosystems……………………………………………………………. 12 4.3. Interaction with metals, and mechanisms of resistance in bacteria……………………… 19 4.3.1. Interaction with lead (Pb (II)), and mechanisms of resistance in bacteria...... 21 4.4. Anthropogenic metal contamination impacts on microbiomes………………………….. 23 4.5. Public health concerns: the link with antibiotics………………………………………… 29 5. Horizontal Gene Transfer: the interconnected community…………………………………… 29 5.1. Definition, process and implications…………………………………………………….. 34 5.2. The gene transfer response facing environmental stressors……………………...... 36 6. Studying bacterial community resilience………………………………………………...... 36 6.1. In-situ vs. in-vitro microbial communities…………………………...………………….. 36 6.2. Deciphering environmental communities: Who is there and what are they doing?...... 6.3. In-vitro conjugation assay to study plasmid permissiveness of a natural microbial 39 community…………………………………………………………………………………… 43 6.4. Study bacterial interactions in bacterial communities…………………………………… 44 7. Aim of the study………………………………………………………………………………. 46 8. Sites of interest: Férin and MetalEurop river sediment community models………………….. 54 9. References…………………………………………………………………………………….. Results Chapter 1: Long-term industrial metal contamination unexpectedly shaped diversity and activity response of sediment microbiome 1. Introduction…………………………………………………………………………………… 70 2. Materials and Methods…………………………………………………………………...... 71 2.1. Sampling, DNA and RNA extraction, cDNA synthesis…………………………………... 71 2.2. High throughput 16S rRNA gene sequencing……………………………………...... 72 2.3. Annotation and generation of the contingency table………………………………...... 72 2.4. Alpha-diversity analysis…………………………………………………………………… 72 2.5. Beta-diversity analysis…………………………………………………………………….. 72 2.6. Identification and validation of functional response groups (FRGs)……………………… 72 3. Results and discussion………………………………………………………………………… 73 3.1. RNA and DNA description of sediment microbiomes…………………………………..... 74 3.2. Deciphering the community tolerance and sensitivity range……………………………… 75 3.3. Significant increase in richness in the contaminated site…………………………………. 76 4. References…………………………………………………………………………………….. 77 5. Supporting Figure File………………………………………………………………………... 80-84 6. Supporting Table File………………………………………………………………….....…… 85-88

Chapter 2: Metal-induced bacterial interactions promote diversity in river-sediment microbiomes 1. Background…………………………………………………………………………………… 90 2. Results………………………………………………………………………………………… 92 2.1. The α-diversity of river-like metal-impacted microcosms remains high over time………. 92 2.2. Community succession in control and metal-impacted microcosms…………………….... 92 2.3. Dynamic interaction network reveals links between keystone species and facilitator bacteria………………………………………………………………………………………. 97 2.4. IncP plasmids and metal-resistant genes content assessed by Q-PCR…………………….. 99 3. Discussion…………………………………………………………………………………….. 100 3.1. The community diversity in metal-impacted river-like microcosms……………………… 100 3.2. Mechanisms of resilience in metal-impacted microbial community……………………… 101 3.3. Microbial interaction: Linking bacteria acting in the public good with the rest of the community………………………………………………………………………………...... 102 4. Conclusion…………………………………………………………………………………….. 103 6. Materials and Methods…………………………………………………………………...... 104 2.1. Sampling methods and microcosm maintenance………………………………………….. 104 2.2. Total and bioavailable metal content in sediments………………………………………... 104 2.3. High-throughput 16S rRNA gene sequencing…………………………………………….. 105 2.4. Measuring the diversity through time……………………………………………………... 107 2.5. Identification and validation of the time response groups………………………………… 108 2.6. Dynamic interaction network……………………………………………………………… 108 2.7. Functional and plasmid-associated gene content assessment by quantitative PCR………. 108 References………………………………………………………….……………………………. 111

Supporting Figure File…………………………………………………………………………... 117-122

Supporting Table File…………………………………………………………………...... 123-127 Chapter 3: Selection and propagation of IncP conjugative plasmids following long-term anthropogenic metal pollution in river sediments 1. Introduction…………………………………………………………………………………… 129 2. Experimental………………………………………………………………………………….. 129 2.1. Sediment sampling and DNA extraction…………………………………………………... 129 2.2. Quantification of IncI, IncF and IncP plasmids by quantitative PCR (qPCR) ……...... 130 2.3. Solid surface filter conjugation assay……………………………………………………... 130 2.4. High throughput 16S rRNA gene amplicon sequencing…………………………………... 130 2.5. Annotation and generation of the contingency table………………………………...... 130 2.6. Statistical analyses…………………………………………………………………...... 131 3. Results ……………………………………………………………...... 131 3.1. qPCR quantification of plasmid oriT and plasmid transfer rates…………………………. 131 3.2. Comparative analysis of SMCs 16S rDNA profiles………………………………………. 131 3.3. Link the permissiveness with FRGs of the communities…………………………………. 132 4.Discussion…………………………………………………………………………………….. 132 4.1. Conjugative plasmids as drivers of SMCs adaptation to long-term metal exposition 132 4.2. Comparative analysis of SMCs permissiveness…………………………………………… 133 4.3. Linking strain permissiveness and activity………………………………………………... 135 4.4. Activity patterns revisited with HGT capacities…………………………………………... 136 5.Conclusions……………………………………………………………………………………. 136 References……………………………………………………………………………………….. 136

Supporting Figure File.………………………………………………………………...... 138-144

Supporting Table File………………………………………………...... 145-150

Chapter 4: Metal drives complex dynamics of IncP-plasmid in two-members bacterial community 1.Introduction……………………………………………………………………………………. 152 2.Materials and Methods…………………………………………………………………...... 153 2.1. Strains, plasmids and growth conditions………………………………………………….. 153 2.2. Assessing plasmid burden……………………………………………………………….... 154 2.3. Conjugation assay in liquid community and cytometry analysis…………………………. 154 2.4. Proteomic analysis by SWATH mass spectrometry…………………………………….... 156 2.5. MRM identification and quantification of the PbrA protein……………………………… 159 3.Results………………………………………………………………...... 159 3.1. Plasmid burden is different across mating species ……………………………………….. 159 3.2. Bi-membered community dynamics assessed by flow-cytometry………………………… 161 3.3. Meta-proteomic profiling of two-members communities…………………………………. 162 4. Discussion…………………………………………………………………………………….. 170 4.1. Lead and mating partner modulated plasmid donor relative fitness………………………. 170 4.2. Propagation dynamics of the conjugative pKJK5 plasmid in metal-impacted environment…………………………………………………………………………………. 172 References……………………………………………………………………………………….. 173

Supporting Method………………………………………………………………………………. 178-180

Supporting Figure File.………………………………………………………………...... 183-189

Heatmap Supporting Information……………………………………………………………….. 190-213 Supporting Table File………………………………………………...... 214-228 Discussion and perspectives 1.Exploring bacterial diversity of metal contaminated sediments and deciphering strategies of resilience……………………………………………………………………………………….. 229 1.1. The diversity of bacterial communities……………………………………………………. 229 1.2. The riverine dendritic and dispersive ecosystem………………………………………….. 230 1.3. The metal-selected anthropogenically-sourced bacteria…………………………………... 231 1.4. Public-good providing bacteria……………………………………………………………. 232 1.5. The particulate and heterogenous river-sediments………………………………………… 233 1.6. Metals as extreme event………………………………………………………………….... 233 1.7. Horizontal Gene Transfer as driver of resilience………………………………………….. 235 2. Is plasmid dispersion a key for the resilience of MetalEurop’s environmental community……………………………………………………………………………………… 236 2.1. The permissiveness of metal-impacted sediment microbial community…………………. 236 2.2. The direct impact of lead on conjugation in-vitro………………...... 237 2.3. The implication of environmental metal-contamination and the dispersive process of antibiotic resistance genes………………………………………………………………... 240 2.4. Plasmid dispersion time-scale for the resilience of metal-impacted microbial communities………………………………………………………………………...... 240 2.5. Going further………………………………………………………………………...... 242 3. Supplementary data………………………………………………………...... 244 4. References…………………………………………………………………………………….. 246 Conclusion………………………………………………………………………………………. 252

VII

Publications, oral communication and poster presentation Publications Jacquiod, S.*, Cyriaque, V.*, Riber, L., Al-soud, W.A., Gillan, D.C., Wattiez, R., and Sørensen, S.J. (2018) Long-term industrial metal contamination unexpectedly shaped diversity and activity response of sediment microbiome. J. Hazard. Mater. 344: 299–307. Pinilla-Redondo, R., Cyriaque, V., Jacquiod, S., Sørensen, S.J., Riber L. (2018). Monitoring plasmid- mediated horizontal gene transfer in microbiomes: recent advances and future perspectives. Plasmid; 99: 56–67. Cyriaque, V.*, Jacquiod, S.*, Riber, L., Al-soud, W.A., Gillan, D.C., Sørensen, S.J, and Wattiez, R. (2020) Selection and propagation of IncP conjugative plasmids following long-term anthropogenic metal pollution in river sediments. J. Hazard. Mater. 382C: 121173 Cyriaque, V. *, Géron, A. *, Billon, G., Gillan, D.C, Nesme, J., Sørensen S.J., Wattiez R., Metal- induced bacterial interactions promote diversity in river-sediment microbiomes, submitted. * Mentioned authors contributed equally to the publication Off-topic Publications Fossépré, M., Trévisan, M. E., Cyriaque, V., Wattiez, R., Beljonne, D., Richeter, S., Surin, M. (2019). Detection of the Enzymatic Cleavage of DNA through Supramolecular Chiral Induction to a Cationic Polythiophene. ACS Applied Bio Materials 2:2125-2136 Delacuvellerie, A., Cyriaque, V., Gobert, S., Benali, S., Wattiez, R. (2019) The plastisphere in marine ecosystem hosts potential specific microbial degraders including Alcanivorax borkumensis as a key player for the low-density polyethylene degradation. J. Hazard. Mater. 380: 120899 Oral communication Cyriaque V., Géron A., Gillan D., Wattiez R., Plasticity of microbial communities in metal-impacted river-like microcosms, 24th NSABS, Ghent, Belgium (2019). Poster presentations Cyriaque V., Jacquiod S., Riber L., Abu Al-Soud W., Gillan D., Sørensen S. J., Wattiez R., "Deciphering strategies of a river-sediment microbial community to cope with anthropogenic metal contamination" in "FEMS19", Glasgow, Scotland (2019) Cyriaque V., Jacquiod S., Riber L., Géron A., Abu Al-Soud W., Gillan D., Sørensen S.J., Wattiez Ruddy, "Deciphering strategies of a river-sediment microbial community to cope with anthropogenic metal contamination" in "BAGECO15", Lisbon, Portugal (2019) Cyriaque V., Géron A., Gillan D., Wattiez R., "Following the impact of metals on river sediments in microcosms: metals as a community manager" in "Symposium of the Belgian Society for Microbiology (BSM), “Microbes in the Spotlight”", Brussels, Belgium (2018) Cyriaque V., Géron A., Gillan D., Wattiez R., "Following the impact of metals on river sediments in microcosms: metals as a community manager" in "17th International Symposium on Microbial Ecology (ISME17)", Leipzig, Germany (2018) Cyriaque V., Géron A., Gillan D., Wattiez R., "Following the impact of metals on river sediments in microcosms: metals as a community manager" in "Ecology of Soil Microorganisms (ESM2018)", Helsinki, Finland (2018) Cyriaque V., Gillan D., Wattiez R., "Metal effect on conjugation frequency: Assessing the exchange dynamics" in "Annual Meeting of the Belgian Society for Microbiology", Brussels, Belgium (2017)

VII

Cyriaque V., Jacquiod S., Riber L., Abu Al-Soud W., Gillan D., Sørensen S. J., Wattiez R., "Adaptive strategies of sediment microbiomes towards long-term anthropogenic metal pollution: An integrative molecular approach" in "2nd international MRM conference ", Ghent, Belgium (2017) Cyriaque V., Jacquiod S., Riber L., Abu Al-Soud W., Gillan D., Sørensen S. J., Wattiez Ruddy, "Adaptive strategies of sediment microbiomes towards long-term anthropogenic metal pollution: An integrative molecular approach" in "BAGECO 14", Aberdeen, Ecosse (2017) Cyriaque V., Jacquiod S., Riber L., Abu Al-Soud W., Milani S., Gillan D., Sørensen S. J., Wattiez R., "Heavy metal accumulation shaped presence and potential activity of sediment bacteria" in "L'industrie du futur (GreenWin)", Mons, Belgique (2017) Cyriaque V., Jacquiod S., Riber L., Abu Al-Soud W., Milani S., Gillan D., Sørensen S. J., Wattiez R., "Heavy metal accumulation shaped presence and potential activity of sediment bacteria" in "Mardi des Chercheurs (MdC2017)", Mons, Belgique (2017) Cyriaque V., Jacquiod S., Riber L., Abu Al-Soud W., Milani S., Gillan D., Sørensen S. J., Wattiez R., "Heavy metal accumulation shaped presence and potential activity of sediment bacteria" in "Annual Meeting of the Belgian Society for Microbiology (BSM) ", Brussels, Belgium (2016) Cyriaque V., Gillan D., Beraud M., Abu Al-Soud W., Billon G., Sørensen S. J., Wattiez R., "Evolution and comparison of metal-contaminated sediment microbial communities" in "ISME 16", Montréal, Canada (2016) Cyriaque V., Gillan D., Billon G., Beraud M., Wattiez R., "Comparison and evolution of the plasmidome in metal-contaminated microbial communities" in "Annual Meeting of the Belgian Society for Microbiology (BSM)", Brussels, Belgium (2015)

VIII

Abbreviations AAOB Autotrophic Ammonia-Oxidizing Bacteria Ab Antibiotic ANOVA Analysis of Variance AOB Ammonium Oxidative Bacteria ARG Antibiotic Resistance Gene ASV Amplicon Sequence Variants BMRG Biocide & Metals Resistance Gene cDNA Complementary Deoxyribonucleic Acid DGGE Denaturing Gradient Gel Electrophoresis DNA Deoxyribonucleic Acid EPS Exopolysaccharide FACS Fluorescence Activated Cell Sorting FDR False Discovery Rate FER Férin FeRB Fe(III)-Reducing Bacteria FISH Fluorescence in-situ Hybridization FSC Forward Scatter FRG Functional Response Group GFP Green Fluorescent Protein GTA Gene Transfer Agent HGT Horizontal Gene Transfer HME-RND Heavy-Metal-Efflux Resistance-Nodulation-cell HSAB Division Hard-Soft Acid Base ICE Integrative Conjugative Element Igeo Geoaccumulation Index IS Insertion Sequence LFSE Ligand Field Stabilization Energy LGT Lateral Gene Transfer LPS Lipopolysaccharide LUMO Lowest Unoccupied Molecular Orbital MET MetalEurop MGE Mobile Genetic Element MNTD Mean Nearest Taxon Distance index MPD Mean Pair-Wise Phylogenetic Distance MPF Mating Pair Formation MPI Metal Pollution Index MRG Metal Resistance Gene nbGLM negative binomial Generalized Linear Model NRI Net Relatedness Index NTI Nearest Taxon Index OriT Origin of Transfer OTU Operational Taxonomic Unit PAH Polycyclic Aromatic Hydrocarbon PBS Phosphate-buffered Saline PCA Principal Component Analysis PCR Polymerase Chain Reaction PEQ Probable Effect Quotient PHB Polyhydroxybutyrate POP Persistent Organic Pollutants

IX qPCR Quantitative Polymerase Chain Reaction QS Quorum Sensing RNA Ribonucleic Acid rRNA ribosomal Ribonucleic Acid SMC Sediment Microbial Community ROS Reactive Oxygen Species SD Standard deviation SDS Sodium Dodecyl Sulphate SEM Standard Error of The Mean SSC Side Scatter SMC Sediment Microbial Community ROS Reactive Oxygen Species SRB Sulphate Reducing Bacteria ssDNA Single-strand Deoxyribonucleic Acid T4SS Type 4 Secretion System TCA Tricarboxylic Acid Cycle TI Toxicity Index TRG Treatment Response Group tRNA transfer Ribonucleic Acid tcDNA total community DNA VGT Vertical Gene Transfer WWTP Wastewater Treatment Plant

General Introduction

1. Foreword Let it be called Earth or Gaia (1), the symbiotic planet is a sink of knowledge which is far from having been fully explored.

Much more abundant on Earth than other organisms (>1030 cells), Bacteria and Archaea are also phylogenetically more diversified than animals and plants leading to more complex interactions. As such, deciphering their ecology processes is presently one of the most challenging tasks (2). Yet, microbial communities play a crucial role in all biogeochemical cycles (3). Emergent properties from joint metabolisms of microbial communities define the health and viability of the biosphere (4). Understanding their dynamics is one of the greatest challenges to respond to current environmental issues (5). In this way, we will be able to fully understand how the biosphere impacts and responds to changing environmental conditions (6). To decipher microbial community dynamics, its organisation, complexity and interconnectivity (3), theories need to be confronted with observations, tested and validated or invalidated (2). The community concept was used first for macroecology. A community is considered to be a multi-species assemblage in which organisms live together in a contiguous environment and interact with each other (6). Gleason and colleagues (1926) added co-occurrent species to this definition (6, 7). In microbiology, however, it is more difficult to delineate the contiguous environment and delimit interactions. These interactions include metabolic interactions, allelopathy, signalling, structural (biofilm formation), trophic level interactions and Horizontal Gene Transfer (HGT). From these interactions, emergent properties may appear, leading the system to respond to external and internal perturbations, such as syntrophy (6) or sequential detoxification (8). The aim of the present work is to decipher the response of a sediment- associated microbial community (Section 2 of this introduction) in a metal-impacted environment.

Anthropogenic metal contamination of soil and sediment is a serious environmental issue. As metals cannot be degraded, they accumulate in the food chain (9–11) and cause non- negligible environmental and health issues (12). Metal contamination could also play a significant role in the emergence, maintenance and proliferation of antibiotic resistance (13, 14) in environmental and clinical microbial communities through HGT. This is particularly alarming considering that anthropogenic metal contamination is several orders of magnitude greater than antibiotic contamination (13). Metals in the environment and their interactions with bacteria are addressed in section 4 of this introduction.

Plasmids (Box 1) are known to be efficient vehicles for HGT into microbial communities, leading to bacterial evolution and adaptation. Nonetheless, it is still unclear how environmental stressors, such as metals, influence the movement of plasmids by conjugation (15–17). Section 5 of this introduction explores knowledges about HGT and the dynamics of plasmid in stressed environments.

The present work focuses on the mechanisms of resilience (i.e., the capacity of a system to absorb disturbance and reorganize itself while undergoing change so as it still retains essentially the same function, structure and identity (18)) of environmental microbial communities facing metal stress. These interactions include intrinsic bacterial capabilities as public-goods and plasmid transfer. The present work delves into these different aspects working on a metal-contaminated river-sediment microbiome in-situ, the Deûle river (Northern France, section 8 of the introduction), in large-scale microcosms, and on bacterial batch co-cultures.

2. The river ecosystem and its sediment microbiome Rivers are lotic ecosystems where watercourses are flowing from an upstream area to downstream sea, lake or river. The global surface area of streams and rivers is estimated at 662,041 km² (19). The top layer of the earth's crust (soil) below the flowing supernatant water is called sediment, defined as “suspended or deposited solids, acting as a main component of a matrix which has been or is susceptible to being transported by water” (20, 21). Sediments are weathering products of rocks and organic material deposited in the riparian ecosystem resulting from a cycle of erosion, suspension and deposition. Sediments are characterised by quality factors (i.e. grain size, mineral and organic content, nutrients, pathogens or contaminants) and by quantity factors (i.e. pH, amount of generated/lost/transported sediments, grain size distribution or channel morphology). Grain size is categorized in clay (<0.004 mm), silt (0.004- 0.06 mm), sand (0.06-2 mm) and gravel (> 2 mm). They determine habitat quality for various benthic organisms (Figure 1) during key life stages providing nutrient and adherent surface with consequences for the rest of the food chain (21, 22). The dynamic river system constitutes a continuous spatial and temporal gradient of physico-chemical conditions from the headwater to estuaries, inducing a gradient of abiotic and biotic succession (23, 24, 25), with the transport of sediment particles downstream (22). Lands also bring soluble components into the river (26), including anthropogenic substances from urban and industrial areas, WWTPs and agricultural area. These sources enrich the watershed with nutrients (27), antibiotics (28), Persistent Organic Pollutants (POPs, (29)) and metals (30, 31).

Depending on the redox potential, the pH, and the organic matter content of sediments, pollutants will adsorb to sediment particles making them a pollutant reservoir (29, 32, 33). Contaminated sediments may thus represent a risk for populations of fish, invertebrates, macrophytes and diatoms (34) because of their capacity of bioaccumulation (35), as well as for river microbial communities and their functional potential (36–39).

Figure 1. The sedimentary bed material hosts a large variety of microorganisms and provides up to 1000 m² of grain surface per m³, serving as matrix for microbial biofilms. Arrows indicate dispersion of cells with the flux of surface water and inside interstitial water. Adapted from Battin et al. (2016) (19). Despite the fact that rivers cover only a small part of the Earth’s surface, they host a disproportional biodiversity (25). Furthermore, they present numerous ecosystem services (benefits that people obtain from ecosystems), such as food supply in the food chain (including fishing), water supply, energy supply, recreation area, navigation, aesthetics (21, 40). They also receive Wastewater Treatment Plant (WWTP) outlets (41). Sediments play a key role in flood protection, water resource maintenance and ecosystem protection (i.e. erosion protection, waste treatment, carbon sequestration and storage, biofiltration, etc.) (21, 40). Sediment microbial communities (SMCs) are key players in biogeochemical cycles, including decomposition or pollutant degradation. They interact with both macroscopic organisms and the abiotic environment (25).

Headwaters Soil River & GW

Litter WasteWater fall Treatment Plant effluent

Leaf River

Estuary Reservoir outflow River WWTP Lotic Sea Lentic River WWTP nutrient pulse

Labile carbon concentration

Figure 2: Exhaustive list of origins of coalescing microbial communities in the riverine ecosystem and soluble components. The brown area represents the mixture of arriving and indigenous communities and environmental components. Arrows indicate the direction of community exchange. GW is for Groundwater. Adapted from Mansour et al. (2018) (25). The river microbial network is established in a directional linear branched structure. These communities are shaped by the riverine flow, which accelerates the dispersal movement of microorganisms from upstream area and from the land (Figure 2), (19, 42, 43). In this way, the river source enriches the downstream communities and maintains biodiversity by meeting local and upstream communities coalescing together (19, 44). This could result in a metacommunity forming a continuum along the river with a constant alpha-diversity, while the beta-diversity is shaped by local habitat resources and interactions (25, 42, 45, 46). The microbiome colonising sediments is dense and phylogenetically diverse (19, 47). Stream and river sediments host 107 to 109 microbial cells per square centimetre (19). The water-saturated part often hosts the main part of the riverine system biomass compared to the pelagic zone (37). Its alpha-diversity was repeatedly higher than in the supernatant water (48, 49), likely because

of their heterogeneous constitution (37). Bacteria are key elements for all biogeochemical cycles, since they are involved in organic matter and other compound degradation and recycling (50), ensuring up to 96 % of total respiration (37).

Taxonomically, depending on the sequencing technique and the studied river sediments, the most represented phyla are Proteobacteria (37, 46, 48–53), Bacteroidetes (37, 48, 50–53), Actinobacteria (46, 48, 49, 51, 53), Acidobacteria (37, 46, 48, 49, 53), Firmicutes (46, 49, 51, 52), Chloroflexi (52, 53) and Planctomycetes (46, 48, 53), with the appearance of other groups, such as Cyanobacteria (48), Nitrospira (37, 52, 53) or Verrucomicrobia (37, 46, 48, 52), with seasonal fluctuations (48).

Most sedimentary bacteria are organised in epipelic biofilms (i.e. sediment-particle associated biofilms, Figure 1) (25, 37, 47) which represent a major part of the river microbial community in comparison with planktonic cells (54). These biofilms are generally very active, intensively occupy the muddy sediments (clay or silt) (46, 55) and process a significant part of the biogeochemical and biodegradation activities (56).

3. Biofilms in river ecosystems Microorganisms often concurrently live in a spatially structured multispecies community attached to a surface and embedded in an extracellular matrix (57–59). These structures, called biofilms, give microorganisms an ideal environment protected from external forces (e.g. flowing water), growth inhibitors (e.g. antibiotics) or predators (57), increasing their residence time. Furthermore, the attachment to a surface which may contain organic molecules, facilitates the contact with the substrate (46). The biofilm is made of the heterogenous stable juxtaposition of microbial cells whose members are dependent on each other’s activities (46, 47, 60). The spatial organisation favours functional coordination, enhancing cooperation in comparison with planktonic cells (46, 61), particularly in multi-species communities whose members interact peculiarly with each other (62).

Biofilm formation starts from planktonic cells adhering to a surface frequently containing organic molecules serving as a substrate. Cells adhere to surfaces through Van der Waals interactions, inducing a reversible adhesion (63). Attached cells then aggregate by cell- cell adherence mechanisms (64–67). The early-stage biofilm proliferates, producing a matrix of exopolysaccharides (EPS; polysaccharides, proteins, and nucleic acids) that form a multi- layered mature organised structure maintained with adhesins, flagella, fimbriae, pili, lipopolysaccharides (LPS), proteins (e.g. SadB, LapA) and DNA. Chemical interactions, such as dipole, hydrogen, ionic, or hydrophobic bonds, may form between the surface and cells (60, 63, 66, 67). The biofilm is formed of juxtaposed cells crossed by water channels that help to distribute oxygen, nutrients, waste and signalling molecules (60, 63, 66). EPS can reach 90 % of the dry mass and play a key role in the biofilm architecture. A biofilm is a reservoir for quorum sensing (QS) autoinducers, extracellular enzymes, and metabolic products, as well as a barrier against pH variations, antibiotics and biocides (60, 66, 68). Under unfavourable conditions (e.g. starvation, lack of oxygen), upper cells detach from the biofilm and turn into planktonic cells, such as free cells can colonise a new physical habitat (64–66).

In aquatic ecosystems, biofilm organisation creates microhabitats with pronounced gradients of dissolved and particulate matter, of oxygen and other electron acceptors (nitrate, manganese, iron, sulphate, and carbon dioxide) (55). Therefore, biofilms are responsible for a large part of the biogeochemical and biodegradation activities (56) compared to planktonic bacterial cells, including the biodegradation of pollutants (57).

Fine sediment epipelic biofilms are mainly composed of bacteria, along with cyanobacteria, archaea, algae, fungi and protozoa. The biofilm taxonomic composition and architecture are dependent on the grain size and roughness, as well as on the zonation profile in depth impacting the light and oxygen dispersion. Generally, the biomass of microorganisms increases with the available area and the roughness of the grain (54). The cohesive matrix, formed by microorganisms closely surrounding and embedding fine sediment particles, clearly stabilises sediments against re-suspension (69). Moreover, the adhesion of molecules to sediment particles makes the first sediment layers a sink of nutrients serving as good substrate for biofilms. The benthic sediment zone is very dense in biomass. This zone hosts autotrophic and heterotrophic organisms (54), including α-, γ- and δ-Proteobacteria, and methanogenic Archaea that play a key role in subsurface water saturated microbial communities (46, 69).

4. Metal contaminated environments and how bacteria deal with them 4.1. Metals, definition and classification Metals are defined as elements which conduct heat and electricity, have a metallic lustre, are malleable and ductile, form basic oxides and are positively charged in aqueous solution. Metalloids possess a variable number of electrons in the d-shell orbitals. They are semiconductor and brittle metals (70, 71). The term “Heavy Metals”, defined according to their high density (from > 3.5 g/cm3 to >7 g/cm3 depending on authors), atomic weight (>23) or atomic number (>20), is often used for hazardous metals and metalloids with highly toxic and ecotoxic properties (71). However, there is no link between the density and toxicity of an element. Another classification is then needed to forecast the interactions between metal ions and ligands and then, assess their toxicity. These interactions will define the speciation of the metal (i.e. different chemical species and proportion of these species in a given niche) that also depends on environmental factors, including pH, organic carbon content, alkalinity, reduction potential, ionic strength and temperature (72, 73). The toxicity of metals depends both on the speciation of the chemical element defining its behaviour as an electron acceptor and its bioavailability (71, 72). Bioavailability is defined here as “the degree and rate of absorption of a substance into a living system, becoming available at a site of physiological activity (74)).

Metals can be classified as Lewis acids considering the reactive vacant orbital or the available lowest unoccupied molecular orbital (LUMO) (71) that define their interaction with available basic ligands (71). The Hard-soft Acid Base theory (HSAB; also known as Pearson’s acid base theory) empirically considers Lewis acids into the Hard, Borderline and Soft categories (Figure 3). Hard acids are non-polarisable and preferably ionically bind to hard basic ligands (e.g. oxygen). Soft acids are polarisable and preferably covalently bind to soft basic ligands (e.g. sulphur and sulphide). Borderline acids form relatively stable complexes with hard and soft ligands, are more difficult to classify and, therefore, it is more difficult to assess their toxicity. Each oxidisation state of a metal must be considered individually in order to classify metallic elements (Figure 3, 72).

Figure 3: Hard-soft acid base (HSAB) classification of metallic ions according to Lemire, et al. (2013) (72). Class A represents Hard metals while Class B represents Soft metals (72). In the biosphere, cadmium is usually found as an oxidised species in the form of Cd (II). It is a soft Lewis acid (Figure 3), mobile at low pH and forming complexes with organic compounds and oxides (75–77). Copper can be found as Cu(I) or Cu(II) that are soft and borderline acids respectively (Figure 3). Copper binds strongly to organic matter and phosphates (75). Lead can be found as Pb(II) and Pb(IV) that are soft and borderline acids respectively (Figure 3). Their mobility is low, and they bind to clay minerals, organic matter and phosphates (75, 78, 79). Zinc is generally found as Zn(II). It is a very mobile borderline Lewis acid (Figure 3) (72). The mobility of these metals makes them relatively bioavailable in the following order Zn>Cd>Pb in sediments (32). It is established that it is the bioavailable fraction of metals that has to be considered to assess the toxicity and risk level of the contamination (74, 80). However, only a few legislative texts consider metal bioavailability. Furthermore, we must take mode of co-action into account as, most of the time, environmental metal pollution encompasses a mixture of several metals. Independent action and concentration addition must then be considered but are difficult to assess as the mode of co-action of most of metals (additive, synergic, antagonist) is far to be deciphered (73).

4.2. Metals in aquatic ecosystems Metal ions are part of primary (i.e. magmatic) rocks that solidify and crystallise on the earth from the cooling down of magma arriving from the earth's mantle through geological processes, such as volcanism and plate tectonics. Soft and borderline metals rarely form their own mineral but substitute hard metals in crystalline structure (e.g. nickel can replace magnesium in the crystal formed of Mg2[Ni]SiO4). Metal ores appear at a later stage of differentiation when metal concentration increases in magma formed of hot residual hydrothermal fluids (81). Metals disperse in the environment through weathering events that disintegrate rocks into stones and particles while erosion tears off particles and metal ions (81). Metals adsorb to soils, leach in waters and end up in groundwater, river and lakes (73) where they disperse in biota and sediment that constitutes a sink for metal ions. Sediments accumulate up to 99 % of metals entering into rivers (11, 73, 82). Metals can be confined in amorphous materials. However, most are soluble or ion-exchangeable, adsorbed to clay or bound to Fe-Mn oxyhydroxides, oxides, organic matter, sulphates and carbonates or lattice in minerals (73, 82). This will define the speciation of sediment-associated metals that may also be classified using five steps of a sequential extraction process (easily extractable and exchangeable, carbonate bound, iron and manganese oxides bound, organic matter bound and residual metals) (82).

The use of metals and metalloids as raw materials, in industries and technologies drastically increased from the industrial revolution and urbanisation of the XVIIIth century (Figure 4; 35, 83) leading to an increasing number and diversity of anthropogenic metal contamination sources including mining activities (84), wood processing (85, 86), shipping (86), dredging, urbanisation and automobile (84), agriculture (soil amendment (87,88) and pesticides (89)) and farming (84, 86), industrial releases (84) and medicine (86). Metals are leached in the environment and end up in groundwater, lakes and rivers resulting in a drastic increase in metal contaminated biota, especially in sediments that constitutes the sink endpoint for metals.

Environmental metal contamination of aquatic ecosystems constitutes a worldwide issue getting more and more attention due to the persistence of metals (they cannot be degraded) (90). Because of their large biomass and their ubiquity, SMCs are the first to be impacted by metals in this peculiar environment.

Figure 4: Evolution of lead (grey), copper (red) and zinc (green) production (solid line) and emission (dot lines) from industrial revolution to 1990. From Järup, L. (2003) (83). 4.3. Interaction with metals and mechanisms of resistance in bacteria Some metals are considered as essential (calcium, cobalt, chromium, copper, iron, potassium, magnesium, manganese, sodium, nickel, selenium, vanadium, molybdenum and zinc) because they are involved in the catalysis of biochemical reactions and electron transfer, the stabilisation of protein structure and the bacterial cell wall. Potassium and sodium are involved in the maintenance of the cell osmotic balance. Transition metals, such as iron, copper and nickel, are used in redox active enzymes involved in respiration, N fixation and photosynthesis. Magnesium and zinc play a role in enzyme, chromatin and DNA stabilisation. Silver, aluminium, gold, lead, and mercury do not have any biological role in the cell and are considered as nonessential (72, 91).

The role of a cation can be inferred from its ability to bind a specific ligand. These interactions are dependent on its HSAB classification (Figure 3, 5), the pH (impacting the pKa of the reaction) and the required Ligand Field Stabilisation Energy (LFSE) depending on its number of d electrons promoting binding geometries (70). As metals can bind to many cellular ligands, an excessive concentration of essential or nonessential metals is toxic. Depending on the preferred ligand interactions, the mechanism of toxicity is different. First, the increased presence of some metals, such as Cr(VI), As(III) and Te(IV), Fe(II) and Cu(II), leads to a metal- induced oxidative stress that is either direct, like for Cu(II), or indirect by catalysing a Fenton

reaction chemistry1. Such a reaction increases the amount of intracellular Reactive Oxygen − • •− Species (ROS; OH , OH , H2O2, O2 ) that may damage DNA and impact enzymatic activities. An increased electron transfer catalysed by iron in aerobic respiration may also lead to ROS production. Other metals, such as Co, Cr(VI), As(III), Te(IV) and Cu(II), may also disrupt Fe- S complexes and other Fe-containing proteins resulting in a release of Fe in the cellular environment and the production of ROS. Covalent bond formed between inactivate cellular thiols and borderline and soft acid metals or metal oxyanions lead to (i) the formation of protein disulphides and (ii) the mobilisation of antioxidant reserves (e.g. glutathione), leaving the microbial cell vulnerable to metal species or ROS (72, 86). The propensity of a metal to provoke an oxidative stress seems to be substrate dependent. For instance, the use of biphenyl instead of succinate by Pseudomonas pseudoalcaligenes KF707 lead to an oxidative stress in the presence of copper by requiring four dioxygenases. The aerobic metabolism generated more ROS (92).

1 2+ 3+ – • Fe + H2O2 → Fe + OH + OH 13

Figure 5: Biological donor ligands classified according to their affinity for hard, borderline and soft metal ions. This affinity is determined by the Electronegativity, oxidation state, polarisability and ionic/ atomic radius of the ligand in accordance with the HSAB theory. Adapted from Lemire et al., (2013) (72).

Metals also have a direct impact on enzymes by catalysing the oxidation of amino-acids decreasing or abolishing the catalytic activity of the active site. The enzymatic alteration is then followed by an increased protein degradation process. Histidine, arginine, lysine and proline are particularly exposed to metal-catalysed oxidation leading to the formation of carbonyl derivatives (72). Metals can also substitute co-factors resulting in a decreased or abolished activity of the enzyme. Nonessential metals bind with greater affinity to thiol-containing groups and oxygen sites than their essential competitors and replace them in their native binding site (91). For instance, dehydratases containing Fe-S clusters, may be inhibited by Ag(I), Hg(II),

Cd(II) and Zn(II) by Fe substitution in a ROS-independent process (72). The displacement of Zn(II) from the active site of the δ-aminolevulinic acid dehydratase by Pb(II), or the displacement of Cu(I) from the active site of a Cu-Zn superoxide dismutase by Ag(I), inhibits the activity of these enzymes. Mononuclear metalloenzymes using Zn(II), Mn(II) or Co(II) as co-factors may bind to Fe(II) in a process also inducing a metal-induced oxidative stress (72, 93, 94). TCA cycle dehydrogenases of Pseudomonas fluorescens appear to be inhibited by metals disrupting the Fe-S cluster by Fe substitution or by oxidative stress induced Fe starvation (95). The aconitase (citrate to isocitrate conversion) was inhibited by Al and Ga (96). The FumA fumarase was inhibited by these metals while its Fe-S cluster free isozyme, FumC overcomes the inhibition (96). In Pseudomonas pseudoalcaligenes KF707, the fumarase activity is inhibited by Cu but the strain does not possess Fe-S free isozymes to compensate the loss of fumarate conversion. Depending on the substrate used as carbon supply, the Cu sensitivity is different (92).

Metals have been reported to disrupt the membrane integrity. Many membrane proteins have active sites for specific metals (e.g. respiration involved proteins or signal transduction). These proteins display a net anionic charge at neutral pH values that attract metal ions (97). This make them the first front line for metal toxicity particularly against histidine, methionine, cysteine, aspartic and glutamic acid residue carrying active sites. By altering the activity of electron transport chains, metals, such as Ag(II), may impact the chemiosmotic transmembrane potential (72, 98). It is known that zinc impacts the activity of Complex I and IV in Pseudomonas fluorescens affecting the ATP-producing pathways (95). By binding to nutrient transporters or acting as a repressor of gene expression, it has been proposed that metals may reduce their uptake activity and provoke cell starvation (72).

The cell membrane is mainly composed of phospholipids including metal-reactive phosphoryl and carboxyl groups attracting metals at neutral and basic pH. The modification in conformation induced by metals could, therefore, disrupt the integrity of the membrane (72,97) and induce its rupture (84). Furthermore, Cu(II) and Cd(II) have been suspected to be involved in lipid peroxidation producing thiobarbituric acid reactive substances that could be lethal.

Finally, toxic doses of several metals, such as Mn(II), Cr(VI), Cd(II), Mo(IV), Zn(II) oxide or Ti(IV) dioxide, have been reported to be mutagenic in bacteria (72, 99, 100), as well as Ag, Cd, Cu, Pb and Zn (101).

To avoid or counter metal toxic effects, bacteria display resistance mechanisms that can be classified into seven categories (Figure 6) that are (i) uptake regulation, (ii) efflux, (iii) extracellular and (iv) intracellular sequestrations, (v) chemical modification, (vi) metabolic by- pass and (vii) repair (72).

Figure 6: Metal resistance mechanisms developed in bacteria. Bacterial cells may either (i, ii) control fluxes of metals (reduced uptake and efflux), (iii) store metal internally (vesicles, production of metal ligand molecules) or (iv) externally (EPS, production of siderophore), (v) detoxify metals through chemical modification (e.g. oxidoreduction process), (vi) shunt metabolic cycles to by-pass metal-impacted enzymes or (vii) repair metal-impacted proteins with chaperones. Adapted from Lemire et al., (2013) (72). Before entering the cell or after being expelled, metal ions can be complexed with extracellular polymers, LPS, proteins and siderophores. In Gram-positive bacteria, carboxyl groups constitute binding sites for metal ions. In Gram-negative bacteria, this is the role of phosphate groups (102). Soluble siderophores regulate metal internalisation and are low- molecular-weight compounds (500–1500 Da), such as mycobactin, ferrochrome or coprogen, originally produced to catch iron in the environment. At low cadmium concentration, the production of pyoverdine (Fe-specific siderophore) and pyochelin (broad specificity siderophore) by Pseudomonas aeruginosa san ai were decreased and an iron storage protein was detected (103). In a context of metal-stress, siderophore can bind other metal ions and maintained out of the cell by down-regulating its intake (72, 104). Pseudomonas aeruginosa 4EA and PAO1 use pyochelin and pyoverdine as siderophore for the sequestration of lead (105, 106). Sulphides produced and externalised by Sulphate-Reducing Bacteria (SRB) participate in

metal ions removal and detoxification (8, 104, 107, 108). Besides, calcium carbonate precipitation by ureolytic organisms can also be considered for metal resistance (109). Accumulation of metal phosphates in LPS membrane of Citrobacter sp. is permitted by an acid phosphatase localised on the outer membrane (110). Pseudomonas aeruginosa san ai was shown to over-produce EPS for cadmium biosorption (103).

The regulation of membrane transporters internalising metal ions can be tightly controlled. Some regulators can detect metals at the femtomole concentration for the down regulation of metal transporters (72). The reduced uptake operates in partnership with efflux pumps that take out excessive metal ions from the cell through specific transporters controlled by sensitive regulators.

In the periplasm, metal-binding proteins can prevent them to pass the inner-membrane (104). Different examples have been highlighted for copper sequestration in the periplasm, such as CusF in E. coli (111), CopK in Cupriavidus metallidurans (112) and CopM in Synechocystis sp PCC6803 (113). Lead sequestration is assured in Providencia vermicola by the metallothionein BmtA that forms metal-sulphites (114) and by PbrB, a undecaprenyl pyrophosphate phosphatase, in Cupriavidus metallidurans. PbrB works in association with the efflux pump PbrA that transports Pb(II) ions in the periplasm where they are complexed with a phosphate preventing them to re-entry in the cytoplasm (79, 115) .

The cytoplasmic membrane carries metal efflux pumps to externalise metal ions including P-type ATPases (72, 91) formed of 6 or 8 transmembrane segments with internal hydrophilic metal translocation sites and involved in ATP binding and hydrolysis (116). P1B- ATPases handle the transport of Cu(I), Ag(I), Cu(II), Zn(II), Cd(II), Pb(II) and Co(II) through the cytoplasmic membrane (104, 116). ZntA, CadA and PbrA transport Pb(II), Zn(II) and Cd(II) (79). Besides, Heavy-Metal-Efflux Resistance-Nodulation-cell Division (HME-RND) are multi-protein structures that coordinate the inner- and outer-membrane to expel metal ions in the outside of the cell. Among them, CzcA exports Co(II), Zn(II) and Cd(II) (104) and CusA exports Cu(I) (117).

To avoid redox and covalent reactions in the cytoplasm, metals can be precipitated with metal oxides, sulphides, phosphates, carbonates, hydroxyl, protein aggregates, or in elemental metal crystals. Cells change their oxidation state and store metals in intracellular particles that are often associated with the cytoplasmic membrane. For instance, lead is often immobilised as lead-phosphate (103). Cytoplasmic proteins, such as bacterioferritin and cysteine-rich

metallothionein or glutathione encoded by genes, such as bmtA or smtA, serve as metal traps inside the cell (72, 101, 103) to avoid direct and ROS effects. Furthermore, the most common mercury resistance systems are lyase and reductase enzymes coded by the mer operon that detoxifies Hg by changing its oxidation state from Hg(II) to Hg0 (91, 118). In Cupriavidus metallidurans, PbrD is responsible of Pb(II) accumulation in the cell (79). Methylation of metallic compounds (As, Sn and Hg) is also used to detoxify them (104). For instance, Rhodopseudomonas palustris possesses arsM conferring As(III) resistance by producing trimethylarsine (119). Elevated synthesis of carbohydrates and proteins was identified as the major process of cadmium elimination in Pseudomonas aeruginosa san ai (103).

The repair of redox-sensitive molecules directly impacted by metals or by ROS can be supported by chaperones or antioxidants (72). The metal-induced oxidative stress can be countered by antioxidant enzymes such as superoxide dismutase that catalyse the dismutation of superoxide (101).

Alternative proteins or metabolic pathways can be used to decrease the dependence of the cell for metal-impacted enzymes (72). In a context of aluminium stress substituting Fe in hemic enzymes, P. fluorescens displays a reduced activity of the aconitase in the TCA cycle and diversifies the downstream enzymes, isocitrate lyase and NADP+ isocitrate dehydrogenase, to maintain the cycle. In order to regenerate NAD+, Pseudomonas fluorescens upregulates the NOX enzyme (120).

Production of exopolysaccharides for biofilm formation plays a role in metal resistance. The EPS matrix includes high molecular weight polyanionic polymers that bind metal cations, such as proteins, humic acids, polysaccharides and nucleic acids (eDNA) (103, 121). These ligands have amino, cyanide, hydroxyl, carbonyl and carboxyl natures leading to an increased metal-biosorption from 20- to 30 times greater in the extra-cellular matrix than in planktonic cells. It delays the metal-impact on cells until the matrix is saturated (121). Sulphurs and phosphates trapped in the matrix induce metal depletion by forming precipitates as grained minerals. EPS have been shown to significantly adsorb Pb(I) cations in Pseudomonas sp. isolated from a hot water spring (122). Therefore, EPS have been proposed for metal bioremoval (123). The sessile mode of life surrounded by EPS could allow a decreased diffusion rates in the matrix and a protection of the cells from metals and ROS (124). The diversification of the biofilm in phenotypically different subpopulations could also permit the metal-resistance of the surviving cells hosted by the biofilm (47).

In the present work (Results-Chapter 4), we used lead as a metal stressor. The following chapter then aims at clarifying interactions between bacteria and lead.

4.3.1. Interaction with lead (Pb(II)) and mechanisms of resistance in bacteria Pb (II) is a non-essential ion and is highly toxic. It changes conformation of nucleic acids and protein, as it has a higher affinity with thiol and oxygen groups than Na(I), Ca(II), Mg(II), Fe(II) or Zn(II). Pb (II) may enter the cell through uptake pathways for essential divalent ions (e.g. Mn(II) or Zn(II)). It may disrupt membrane functions, oxidative phosphorylation and the osmotic balance (79, 84) and was shown to reduce growth and the cell size of Pseudomonas aeruginosa 4EA (105). To defend itself against Pb(II), P. aeruginosa 4EA massively produces pyochelin and pyoverdine siderophores (105). Cupriavidus metallidurans CH34, used as a model for metal resistance mechanisms, displays a pMOL30-encoded pbrUTRABCD operon. The operon is bi-directional (Figure 7a). In one direction, PbrR and PbrT are constitutively transcribed. PbrU is not expressed in C. metallidurans CH34 and could encode a permease. PbrR is a MerR-like regulator controlling transcription of pbrABCD genes. When PbrR binds a Pb(II) cation, it up-regulates the transcription of PbrA, PbrBC and PbrD (125). PbrT is a lead uptake permease (79). PbrD is a Pb(II) sequestration protein reducing its toxic effect in the cytoplasm. PbrA is a PIB-type ATPase exporting Pb(II) in the periplasm (79, 125) and is also permissive to Zn(II) and Cd(II). PbrB and C form 1 gene in C. metallidurans CH34 coding for an undecaprenyl pyrophosphate (C55-PP) phosphatase and a lipoprotein signal peptidase, respectively. PbrB removes the β-phosphate from pyrophosphate. The phosphate group is freed in the periplasm where it binds Pb (II) exported ions (Figure 7b) resulting in the sequestration of lead-phosphate salts in the periplasm. Zn(II) and Cd(II) associated phosphates are soluble in water impeding sequestration of those ions (115). The PbrA efflux pump can also be assisted by other transporters, such as ZntA or CadA for the efflux of Pb(II), Zn(II) and Cd(II), whose primary role is zinc homeostasis (79, Figure 7b). The pbr operon was found on the plasmid of several bacteria with different configurations, such as Alcaligenes faecalis, Shewanella frigidimarina, Klebsiella pneumoniae, Herminiimonas arsenicoxydans or Ralstonia picketti. CadA and ZntA efflux pumps were found in other bacteria, such as Pseudomonas putida KT2440 (CadA2), S.aureus (CadA) and E.coli (ZntA) (79).

(a)

(b)

Figure 7: Lead resistance mechanism in Cupriavidus metallidurans CH34 as encoded on the pbr operon (a) involving PbrD, PbrA and PbrBC proteins for the detoxification, efflux and sequestration as a phosphate salt. The PbrA efflux system is helped by CadA and ZntA efflux pumps (b). Figures were obtained from Jarosławiecka and Piotrowska-Seget, 2014 (79). OM: Outer membrane; IM inner membrane. Other mechanisms may be set up to insure lead resistance. In response to lead elevation, C. metallidurans CH34 displays an increased transcription a fecR regulator inducing iron storage protein (ferritin), FeS cluster biosynthesis (involved in the mobilization of Fe and S atoms from storage) and a putative ferredoxin (125). Glutathione metabolism-involved proteins (glutathione S-transferase, glutathione reductase and γ-glutamyl transferase) were up-regulated as well as proteins involved in polyhydroxybutyrate (PHB) formation (125).

The first barrier against metal ions, including Pb(II), is the cell wall. In Gram-negative bacteria, LPS bind cations and in Gram-positive cells, peptidoglycans and teichoic and teichuronic acids insure cation sequestration. EPS production is also involved in extracellular lead sequestration (79). The high content in uronic acids in EPS produced by Paenibacillus jamilae make them highly Pb(II) specific (126).

Precipitation of Pb(II) outside or inside the cell can be achieved by forming phosphate salts, in various forms of salts (e.g. Pb3(PO4)2, in S. aureus, Pb9(PO4)6 in Vibrio harveyi). Pb(II)- binding zinc-homeostasis involved metallothionein were also found in Pseudomonas aeruginosa WI-1(100) as well as a super-oxide dismutase in Streptomyces subrutilis (127).

4.4. Anthropogenic metal contamination impacts on microbiomes Disturbances, such as excessive metal levels, can change significantly the community structure, particularly in riverine ecosystems (42). Contrasting results have been obtained and the impact of metals on the diversity of microbial communities may be negative or positive (128). A decreased diversity was assessed in copper impacted field soil (129) and activated sludge (130), in Pb/Zn ore impacted soils (131), in Au–Ag mine tailings (132), in acid wastewater impacted river sediments (133) or in soils surrounding smelter with a decrease in biomass and phosphatase activity (134). On the opposite, Gillan and colleagues found no change in biomass and activity of the microbial community in an 80-years metal contaminated fjord but revealed community structure variations via Denaturing Gradient Gel Electrophoresis (DGGE). That revealed a long-term adaptation and functional recovery over time (135). Xu and colleagues (2017) found no impacts of Zn and Pb on the soil microbial community richness but highlighted modifications in the community structure (136). Similarly, Sutcliffe and colleagues showed significant impact of uranium on the metabolic potential of mine microbial communities but no significant impact on their richness (137). In the same way, no significant effect of metal bioavailability on Operational Taxonomic Unit (OTU) richness was found in a copper contaminated grassland soil (138). However, the copper impact was demonstrated subsequently on the microbial diversity of the active part of this community by confronting the anthropogenic metal contamination gradient to the 16S rRNA obtained from RNA extracts (85). Other authors nuanced the analysis showing that the impact on diversity was dependent on the metal and its concentration (139, 140). Authors then agree that metal deeply disturb the community structure leading to species selection and replacements. It seems that metal effects on a microbial community are dependent on the sensitivity of its microorganisms and on the factors that contribute to the toxicity of these metals, such as the size of particles in soil, pH, Eh, clay content or iron oxide content that can also be modulated by the evolution of the structure the microbial community itself. Furthermore, the response would be different when assessing a direct acute toxicity induced by metal loading or a long-term effect of chronical metal contamination (128).

In river sediments, both types of results were obtained, revealing insignificant or negative impact of metals on the alpha-diversity of river sediment microbial communities. Metals were shown to decrease the diversity of the Cauvery River (India) sedimentary microbial community (141) or in the Yellow River (China) (133) but had no apparent effect on the diversity of the community extracted from Xiangjiang River (142) or on Nanfei River (China) (52) sediments. Nonetheless, the microbial community structure (beta-diversity) in river

sediment is inevitably impacted by the presence of metals. They favour Proteobacteria (133, 143) especially α-, β- Proteobacteria, (133) and δ-Proteobacteria (143, 144), Bacteroidetes (133, 141, 143), Arthrobacter (141), Acidobacteria (142), Actinobacteria (52), Firmicutes (133, 145) (especially Clostridiales (52)) or Anaerolineaceae (52, 144) compared to control uncontaminated area.

Metals impact communities by selecting or deselecting bacterial metabolisms with a differential effect depending on the type of metal, its concentration and the sensitivity of the targeted species. For instance, aerobic respiration is globally negatively impacted by metals (146). However, while Cd(II) ions replace Fe(II) and Cu(II) in metallo-enzymes involved in aerobic denitrification process in Pseudomonas aeruginosa san ai, at low concentration, the overproduction of these enzymes maintained the metabolic activity (103).

Some anaerobic digestion and respiration metabolisms are also impacted. At low concentration, metals stimulate methane production (147) but acidogenesis (148) or methanogenesis (e.g. Archaea of the group Methanomicrobia) (137, 148, 149) are negatively impacted by high metal concentrations. Besides, hydrogen and acetate conversion together with sulphate reduction by Sulphate Reducing Bacteria (SRB), are thermodynamically advantaged over methanogenesis (Equations 1-4; 147, 149) :

Sulphate reduction

2− + − 4퐻2 + 푆푂4 + 퐻 → 퐻푆 + 4퐻2푂 ΔG0’=-151.9 kJ

− 2− − − 퐶퐻3퐶푂푂 + 푆푂4 → 퐻푆 + 2퐻퐶푂3 ΔG0’=-47.6 kJ Methanogenesis

− + 4퐻2 + 퐻퐶푂3 + 퐻 → 퐶퐻4 + 3퐻2푂 ΔG0’=-135.6 kJ − − 퐶퐻3퐶푂푂 + 퐻2푂 → 퐶퐻4 + 퐻퐶푂3 ΔG0’=-31 kJ Where Gibbs free energies at 25 °C were calculated at standard conditions (i.e. solute concentrations of 1 M and gas partial pressure of 105 Pa).

In addition, SRBs display a higher affinity for H2 than methanogens and in certain conditions, produced sulphides can be more toxic for methanogens than SRB. In addition,

metal-resistant SRBs seem to be resistant to higher metal concentrations and can then outcompete methanogens. However, sulphide complexation with metals could induce the recovery of methanogens (147).

Metals also have a negative impact on Fe reduction by Fe(III)-reducing bacteria (FeRB) but some, such as Geobacter related species, developed metal tolerance. Furthermore, other members of the community, such as Clostridia and Sedimentibacter, able to sequester metals, or SRBs Desulfovibrio and Desulfibacterium, may permit the entire community to thrive (150).

Metals, especially Cu, As, Cd, Hg, U and Pb, inhibit global nitrification processes leading to a decreased respiration rate (137, 151–153). However, if copper and zinc showed a negative impact on nitrification, it seems that iron had a positive impact on the activity of anaerobic ammonium-oxidising species (154). Metals impact on enzyme activity for nitrogen- fixation and denitrification is metal and concentration dependent. For instance, aluminium and copper negatively impact the nitrogenase, nitrate reductase and nitrite reductase enzymatic activities while iron and molybdenum induced these enzymes up to a certain threshold and then negatively impact the enzymes beyond it (155). When investigating the metagenome of uranium impacted sediments, deselection of genes involved in nitrogen fixation depended on uranium concentration probably following the changes in the microbial community structure (137).

4.5. Public health concern: link with antibiotics Since they cannot be degraded, metals accumulate in soils and sediments and pose biotoxicity and bioaccumulation issues, affecting both abiotic conditions and biological activity. They exert a selective pressure on a long time-period (156), enter the food chain (11) to finally affect human health (84).

Recently, concerns emerged with regard to the selection of antibiotic resistance in metal contaminated ecosystems (13, 157). Antibiotics are any molecule (natural or synthetic) that inhibits the growth and/or kills microorganisms. They are produced naturally by bacteria and fungi which carry resistance genes for the antibiotic they produce (157). Other bacteria exposed to antibiotics can acquire antibiotic resistance through mutation of targeted enzymes and porins (for the reduction of the influx) or the acquisition of antibiotic resistance genes (ARGs) for drug sequestration, drug inactivation and rapid efflux, from resistant bacteria via HGT (see section 5) (13, 157). Mechanisms of resistance for antibiotics are similar to those used for metal resistance (13).

Antibiotics have probably been used unconsciously by humans since the antiquity, during the “pre-antibiotic era”, as tetracycline was found in a human skeletal dating from 350– 550 BCE in the ancient Sudanese Nubia (158). At the beginning of the 20’s century, synthetic molecules were screened and 6 of them gave promising results for the treatment of syphilis. In 1928, penicillin was discovered by Alexander Fleming and replaced earlier drugs in the 40’s to toggle in the “antibiotic era”. From 1970’s, new advances were mainly achieved by modification of existing antibiotics (158). Antibiotics are used to treat or prevent bacterial infections for humans, swine, livestock and poultry (159) and promote food animal growth (160). Their intensive use and massive production lead to environmental contamination (161– 164), the selection of antibiotic resistant bacteria and Antibiotic Resistant Genes (ARGs) (e.g. 153, 157, 158). Antibiotic resistant bacteria became an important health issue because the antibiotic resistome (aka entire set of ARGs) spread throughout clinical and environmental microbiomes (165, 166). Bacteria, such as Staphylococcus aureus, Escherichia coli, or Salmonella species (167), also tend to carry Metal Resistance Genes (MRGs) co-located with ARGs on the same plasmids (Box 1) (157).

Combined contamination by metals and antibiotics are observed in hospital effluents, in urbanised areas, particularly WWTPs (168, 169) and in agricultural environments (156, 170) leading to a co-selection process. Due to the similar structural and functional characteristics of both systems, some genes can confer resistance to both antibiotics and metal (cross-resistance) leading to physiological cross-selection along with genetical cross-regulation leading to co- expression (13, 171, 172). Furthermore, the co-occurrence of ARGs and MRGs in genomes and Mobile Genetic Elements (MGEs; see Chapter 5), such as large plasmids, accelerates co- selection (13, 157, 171, 173) even at sublethal concentrations (170) and even when metals are the unique contaminant (Figure 8). A proportion of 17% of bacterial genomes, and 5% of plasmids, carry both ARGs and Biocide & Metals Resistance Genes (BMRGs). About 35% (227/ (420+227)) of the BMRG carrying plasmids also carry ARGs (Figure 9).

As far as we know, co-occurrence of ARGs and MRGs on plasmids mainly concern mercury resistance but resistance genes for arsenic, nickel, antimony, cobalt, iron, copper, cadmium, zinc and silver in genomes were connected to many ARGs. Much co-selection processes are, then, likely insured by chromosomal BMRGs and plasmid-borne ARGs revealing the restrained chance of biocides and metals to directly promote HGT. However, the selection pressure due to biocides and metals might mobilise resistance genes of plasmids carried by metal resistant bacteria. Besides, plasmids with co-located BMRGs and ARGs are more likely

conjugative and carry toxin/anti toxin systems that favours their diffusion and maintenance in the community (157).

In this context, understanding the dynamics of conjugative plasmids in metal impacted environments, such as rivers, is of the greatest importance, especially as we know that anthropogenic metal contamination is several orders of magnitude greater than antibiotic contamination (12).

Figure 8: Link between antibiotic and metal resistance in the metabolism and genome of the bacterial cell involving cross-resistance via proteins involved in both antibiotic and metal resistance; co-regulation and co-expression involving proteins whose activity or transduction/translation are regulated by the same ligands (antibiotic or metals); and co- resistance and co-transfer of ARGs and MRGs co-located on an MGE as plasmids. Adapted from Pal et al., 2017 (171).

Figure 9: Resistance information in genomes (2,522 genomes, left) and plasmids (4,582 plasmids, right), carrying either ARGs, BMRGs or both. Obtained from Pal, et al., 2015 (157).

5. Horizontal Gene Transfer: the interconnected community 5.1. Definition, process and implications To spread in a bacterial community, genes use two paths: vertical and horizontal transfer. Vertical Gene Transfer (VGT) is the inheritance of a gene from a parent cell to the daughter cells by cellular division (174). Horizontal (or Lateral) Gene Transfer (HGT or LGT) is the share of genetic material between two cells with no parent–offspring relationship (175). Different mechanisms are known to transfer genes horizontally: (i) Transformation is the uptake of free exogenous DNA from the environment, (ii) cell fusion insures bidirectional gene transfer between cells, (iii) Gene Transfer Agents (GTAs) allow the delivery of small random pieces of host genome in capsids, (iv) transduction is the acquisition of DNA fragments through phage predation and (v) Conjugation is the transfer of genetic material requiring cell contact between the donor and recipient cells (175).

Bacterial “sexuality”, today known Box 1. Bacterial plasmids are extrachromosomal DNA elements capable as conjugation, was first described in the of autonomous replication. They may recruit 1940’s in Escherichia coli. Joshua the host-cell machinery for a part of the Lederberg demonstrated that the mix of two process and for transcription and translation. They carry no essential gene for the host, but auxotrophic bacteria for different amino- essential genes for their replication, acids and vitamins turned them in partitioning and transfer (i.e. conjugation) and non-essential accessory genes encoding prototrophic bacteria (i.e. able to synthesise beneficial traits. Accessory genes can confer all the compounds needed for growth) (176, virulence, resistance to toxins (e.g. ARGs 177). Today, “conjugation” is the transfer of and MRGs), and metabolic functions (e.g. nitrogen fixation in rhizobia). Genes are 2 Integrative Conjugative Elements (ICEs) or organised in discreet operons forming a plasmids (Box 1) by cell contact (178). J. patchwork of gene clusters as a consequence of genetic recombination. Plasmids may also Lederberg proposed the term “plasmid” in be selfish entities and maintain in the cell by 1952 for “extra-chromosomal hereditary different mechanisms (discussed in the main determinants” (179). It is now defined as text) and/or by drastically decrease their cost by co-adaptation with the host or by small (0.85 to over 1000 kbp) circular or delegating essential functions (as linear (178, 180, 181) self-replicating MGEs conjugation) to other co-occurring conjugative plasmids (‘mobilisable’ of double-stranded DNA that do not carry plasmids) (192, 193) . any essential genes. However, they often

2 “Self-transmissible Mobile Genetic Element (MGEs) that encode the machinery for conjugation as well as intricate regulatory systems to control their excision from the chromosome” (262) 27

carry useful genes, such as ARGs, MRGs, and virulence factors, or can confer metabolic abilities (182) compensating the burden plasmids impose to their host (181, 183). Plasmid- borne genes can recombine with the host chromosome. Conjugation is one of the most important HGT processes, as it permits DNA transfer between remotely related species (178). Therefore, conjugation may lead to the fast adaptation to new environments by dispersing beneficial genes in a community.

Conjugation for the transfer of a plasmid from a donor to a recipient cell involves plasmid genes that permit (i) the membrane associated mating pair formation (MPF genes) and (ii) the mobilisation of the plasmid (MOB genes). The pilus/adhesine assembled by the transferosome (type 4 secretion system, T4SS) comes into contact with the recipient cell (184, 185). One strand of the plasmid DNA is recognised at the origin of transfer (oriT) by the relaxase, phosphodiesterase subunit of the relaxosome. The relaxase unwinds plasmid dsDNA, cleaving the specific strand and initiates rolling-circle type replication (184, 185). The relaxase associates with a coupling factor (T4CP) that, in turn, is recognised by the exporter (T4SS). The transfer of ssDNA is led by the relaxase through the T4SS in the 5’-3’ direction, and the plasmid is circularised in the cytoplasm of the transconjugant. Finally, the replication of ssDNAs is initiated in the transconjugant cell followed by supercoiling (Figure 10) (186). Stable transconjugants are insured by ssDNA binding, anti-restriction and SOS inhibition proteins encoded in the leading region of the plasmid (187). Relaxase-catalysed reaction is mechanistically similar in both Gram-positive and Gram-negative bacteria but Gram-positive relaxosome does seem to require auxiliary proteins to insure a good conformation of the plasmid DNA for cleavage by the relaxase. The T4SS system of Gram-positives is simpler as only one membrane has to be crossed (188). Plasmids can spread between phylogenetically close bacteria (narrow-host range plasmids) or between phylogenetically diverse bacteria (broad-host range plasmids). Conjugation between Gram-positives and Gram-negatives have even been detected (189) as well as between the three domains of life, Bacteria, Eukarya, and Archaea (174). Bacteria use a restriction-modification system mobilising a restriction endonuclease recognising a palindromic tetramer or hexamer to identify foreign DNA as their own is methylated on those palindromic sites (by modification enzymes). Bacteria sharing similar restriction-modification proteins are more likely to exchange DNA and dupe the system. Narrow-host range plasmids that share a long-time history with its host acquire similar signature (190). Broad-host range plasmids are either small (because less likely to contain a sequence recognized by the modification/restriction system) or, such as IncP plasmids, remove restriction

target sites from their sequence. It has been shown that very large spectrum of hosts can be invaded by broad host range plasmids (191).

Plasmids are conjugative if they carry the suite of genes for their own transfer. Non- conjugative plasmids can be mobilisable by the machinery of a conjugative plasmid if they are not too large and at least encode a relaxase that recognise an oriT sequence in cis (Figure 11; 178).

Figure 10: Conjugation process between a donor bacterial cell and a recipient cell involving pilus attachment and retraction, DNA transfer and establishment of the plasmid in the transconjugant cell. Obtained from Low et al. (2001) (186).

Figure 11: On plasmid type: (A) Distribution of plasmid types (conjugative, mobilisable or non-transmissible) according to the size (kb). (B) Genetic constitution of transmissible mobilisable or conjugative plasmids. The condition for being mobilisable is to carry an origin of transfer (OriT) and encoding a relaxase. The condition for being conjugative is to encode a T4SS system. (C) Interaction in the process of conjugation between a conjugative plasmid and a mobilisable plasmid. The relaxase cleaves a specific site within oriT, starting conjugation. Obtained from Smillie et al., (2010) (178). Plasmids are classified in incompatibility (Inc) groups in which members cannot co- exist in the same host-cell if they display either a similar replication (rep genes) or partitioning system (par genes) (181, 194) avoiding segregational incompatibility (i.e., cell division leading to a daughter cell missing one type of the rival plasmids) (195). The exclusion can be symmetric (plasmids exclude each other) or directional (one plasmid outcompetes another) (195). Their co-existence can be prevented by surface and entry exclusion mechanisms that impede the cell- to-cell contact formation or attachment of the T4SS system to the surface of the recipient cell and is, in most cases, coded by the T4SS operon itself (181). Therefore, closely-related plasmids are most likely incompatible, while distantly-related plasmids tend to be compatible. Classification is made by letter (e.g. IncF, IncG, IncH, IncI, IncL, IncM, IncN, IncP, IncU, IncW or IncX). Some, such as IncF, IncI, IncN or, IncX are hosted by a narrow range of host (here, Enterobacteriaceae; 181) and others, such as IncP plasmids, are broad-host range plasmid and can spread in a diverse phylogenetic range of hosts. Incompatibility groups could be an essential factor limiting plasmid spread in a bacterial community (196).

Plasmids impose a metabolic burden to the host cell (Figure 12). The receipt of the plasmid induces a SOS response leading to an increasing number of ssDNA. It induces mutations and recombination and reduces the bacterial cell division. Plasmid genes that are primary expressed inhibit the SOS response (197). When the plasmid settles in the host cell and replicates, the rep genes encoded by the plasmid recruit chromosomal proteins that may lead to a decreased chromosomal replication and the engagement of the SOS response. The conjugation is energetically expensive (Figure 12) as it mobilises plasmids genes for ATPases, the synthesis

of the T4SS and the escort plasmid DNA (184, 188). To diminish this cost, plasmids regulate their transfer. Transfer genes (tra) are constitutively off and are activated by external factors (Figure 14; 184). Moreover, the expression of plasmid genes (transcription and mainly translation) needs energy (Figure 12). The different GC content of plasmid DNA changes the usual codon usage of the cell and may lead to ribosome sequestration, tRNA and amino-acid starvation. This could induce a decrease of translation efficiency, mistranslation and misfolding. Finally, the interaction with other MGEs might lead to SOS response and deleterious effects for the cell. (197).

Figure 12: Plasmids impose fitness costs to the transconjugant cell. (i) The conjugation process requires ATP consumption. (ii) The T4SS system expressed during mating may make the cell sensitive to T4SS- specific phages. (iii) Recognition of exogenous DNA sequence lead to the up-regulation of RecA nucleoproteins that trigger the SOS response. (iv) Furthermore, the overexpression of plasmid replication proteins (rep) can lead to sequestration of the cellular DNA replication machinery, stopping chromosomal replication, inducing the SOS response, and inhibiting cell division. (v) The expression of plasmid encoded genes also imposes a burden to the host (transcription, translation, and interactions between plasmid- and chromosome-encoded proteins). AT-rich genes can induce sequestration of RNA polymerases and different codon-usage can induce tRNA/amino-acid depletion. To decrease these costs, expression of plasmid-encoded genes can be regulated by histone-like nucleoid structuring proteins (H- NS). (vi) The integration of plasmid genes in the host chromosome may also produce deleterious effects associated with the disruption of host genes. (vii) Finally, the co-occurrence of several plasmid can increase mediated costs. Obtained from San Millan, & MacLean (2017) (197).

The metabolic burden imposed by plasmid maintenance decreases the fitness of the host and could lead to plasmid loss if that fitness drop is not balanced by advantageous traits conferred by carried adaptive genes. When a plasmid fails to be maintained in the daughter cell, its lineage is “cured”. To avoid being lost in the microbial community, plasmids promote their stability by different mechanisms coded by its genes (181). Ensuring a high copy number in the cell guarantees plasmid copies to both daughter cells even by random segregation. To facilitate the process, the multimer resolution system (mrs) encodes for a resolvase enzyme (res) that untie oligomers of accidentally recombined plasmids into monomers (183). Low copy number conjugative plasmids (< 10 copies/cell) use an active segregational complex (the segregosome, par genes) to avoid segregational loss in one daughter cell. The segregosome is composed of an ATPase (producing energy) and a DNA binding protein that attach to a plasmid centromere region. This partition systems also define plasmid incompatibility (181, 183). Another strategy consists in stopping growth or kill plasmid cured cells with a post-segregational killing (psk) system making cells addict to its plasmid. Plasmid DNA encodes for a stable toxin (bactericidal or bacteriostatic) and an unstable antitoxin (a protein counteracting the toxin or a small RNA controlling toxin production) (181, 198). Finally, plasmids can re-infect cured cells and a sufficient transfer rate could insure its stability in the community (181, 183).

The mobilome (aka the pool of genes located on MGEs of a community) represents the communal pool available to all the permissive members. Plasmids are vessels for the transportation of other MGEs such as transposons, integrons and mobile gene cassettes (Figure 13). They are able to translocate between DNA supports thanks to different enzymes. Recombinase and transposase insure, respectively, homologous recombination and movement and insertion of transposons. These elements can be translocated from a chromosome region to a plasmid (and vice versa) and disperse in the permissive bacteria of the community via plasmids (181). MGEs also transfer through phages (transduction). As they carry a lot of beneficial accessory genes caught from chromosomes, the communal pool then becomes a very dynamic reserve of useful accessory gene and a powerful tool of adaptation.

Figure 13: Imbrication of MGEs. Site-specific recombination incorporates gene cassettes into integrons by an integrase. Integrons insert into transposons (mobile gene islands flanked by insertion sequences (IS)). Transposons can either be inserted in a chromosome or a plasmid by a transposase if those DNA matrices contain the same IS. The plasmid is then a vessel for the transportation of genetic information. Obtained from Norman et al., 2009 (181). Because it samples the adaptative accessory gene pool from the community (199), HGT permits bacteria to either invade a new ecological niche or improve its performances (forms a new ecotype). They bring new capabilities to the host, such as catabolic pathways, virulence factors, and antibiotic or metal resistances (200), for survival under local selective pressure (199). Known plasmid metal resistance genes concern Zn, Cd, As, Ag, Hg, Cd, Cu or multidrug resistance (157, 201–203). IncP-1 plasmids are broad-host range plasmids with the potential to transfer to a very diverse pool of bacteria and can, therefore, directly connect large proportions of an ecosystem (41, 191). They were previously identified or found as predominant in isolated plasmids from agricultural soils and WWTPs (204–208). They are also known to carry catabolic and antibiotics/metal resistance genes (205, 209).

5.2. The plasmid transfer response facing environmental stressors To understand the dynamics and the role of plasmids in the community resilience, one must understand the impact of environmental parameters on the permissiveness of the community (aka ability to receive a plasmid in terms of transfer frequency and phylogenetic diversity of its transconjugants) (16, 41). As mentioned before, the conjugation machinery is constitutively off. Factors, such as surrounding recipients, nutrients, cell density or environmental signals, can induce tra genes, activating the conjugation process (competent cell) in a part of the population (Figure 14). In doing so, the community avoid exposing all its members to plasmid-associated burden and to pilus-specific bacteriophages (Figure 12) (185). Environmental cues affecting plasmid settlement in a community are from abiotic (e.g. temperature, pH, moisture content, nutrient availability, presence of surfaces and texture, O2, toxic compounds) or biotic (e.g. plant roots, grazing, predatory) nature (199, 210, 211). Nutrients are a limiting factor as they promote bacterial biomass and activity (199, 211) making the rhizosphere, or manured soils, hot spots for HGT (210). Extreme temperatures decrease the bacterial competency (212) by either decreasing bacterial activity (low temperature, (213) or impacting bacterial modification-restriction system in heat shock stress (214). Toxic compounds have been shown to modulate plasmid dispersion either negatively or positively. Ionic liquids (may be due to increased cell membrane permeability; 215), SDS (possibly by repressing the restriction-modification system of the host; 216), antibiotics (at sublethal concentration, antibiotics may serve as signalling molecule inducing conjugative transfer or competency; 15, 217) have all been shown to increase plasmid transfer. Zinc was shown to select for czc genes-carrying plasmid transconjugant in zinc polluted soils (218). Long-term manure exposure also promoted permissiveness (219). On the contrary, short-term addition of herbicides and fertilisers had a negative impact on transformation (213). A study using a broader panel of metals has shown their negative impact on the permissiveness of a soil- associated microbial community with significant differences depending on taxonomy, metal and metal concentration (16).

Figure 14: Positive regulation tra genes (gene encoding for the transfer of plasmids and ICEs by environmental factors followed by a negative feedback loop once Dtr complex and the T4S apparatus are formed. Obtained from Koraimann and Wagner, 2014 (185). Culture-independent techniques have shown, in undisturbed soil environments (191), in manure-treated soil (219) or in WWTP (41, 220) that plasmids can spread in a large phylogenetic spectrum of hosts resulting in different transconjugant pools depending on used plasmid and plasmid donor (41, 191, 219, 220). These studies also indicate key species that could serve as a plasmid transfer hub such an Arcobacter correlated with Aeromonas veroni (191) or the super-permissive OTUs from Enterobacteriales, Burkholderiales and Staphylococci (41). The culture-independent approach has also permitted to assess the direct impact of environmental factors, such as metal-stress. Klümper and colleagues have documented a metal-associated decrease in plasmid transfer frequencies in a soil-associated bacterial community in a dose-dependent manner with no impacts on the transconjugal pool richness. The authors suggested that the reduced permissiveness could be due to a decreased metabolic status and/or a direct effect of metals on plasmid transfer, replication or expression. They suggest a probable correlation between the permissiveness of OTUs and their phylogeny suggesting that metals could regulate the CRIPR-Cas or restriction-modification system. Still, the stress-induced modulation is unique for each OTU (16).

6. Studying bacterial community resilience 6.1. In-situ vs. in-vitro studies The approach for studying biological mechanisms in environmental and applied microbiology uses in-situ or/and in-vitro systems to highlight processes, understand them, and verify emerging ecological theories (2). Both approaches must be used complementarily, as in- vitro studies bring more amenability and tractability, while in-situ studies allow us to observe more complex and relevant phenomena (Figure 15). In-situ observation can be confirmed in microcosms, and theories can be tested in in-vitro synthetic communities. Inferred theories can then be verified in a larger panel of in-situ studies and applied to understand and manage ecosystems (221, 222) or for microbial engineering (223–225).

Figure 15: Different approaches to studying microbial communities from the most natural but complex environments to simplified and controlled in-vitro approaches. Obtained from Ponomarova et al., 2015 (226). 6.2. Deciphering environmental communities: Who is there and what are they doing? Deciphering functions and taxonomical composition of in-situ microbiomes by “omics” approaches allows a broad part of the community to be covered, including uncultivable organisms, as the vast majority of the microbial diversity comes from unculturable microbes. Direct analysis through metagenomics, metatranscriptomics or metaproteomics also avoids biases imposed by laboratory cultivation (227). “Omics” approaches allow the genetic potential of the microbiome to be disclosed by analysing its DNA signature, focusing on the active population using its RNA signature (227), or assessing the functional activity of the community by assessing its proteomic profile. Quantitative PCR on DNA or cDNA (RNA) extracts, or Multiple Reaction Monitoring on protein extract, permit targeted genes, transcripts or proteins to be followed in a community. The diversity of PCR amplified DNA or cDNA fragments can be obtained by amplicon sequencing to assess the diversity of a functional gene or the

taxonomic diversity by analysing the 16S rRNA profile of the community. On the contrary to direct metagenomic sequencing, this technique introduces biases through the PCR amplification (differential primer efficiency for the different sequences) and the amplicon sequencing (differential discriminatory power of the targeted region for the different taxa) (228). However, amplicon sequencing is cheap, fast, requires less bioinformatic cleaning steps and brings a good resolution compared with metagenomics-based 16S rRNA profiling. 16S rRNA amplicon sequencing remains a trustful method to assess taxonomical trends when comparing different conditions and/or the temporal dynamics of a microbial community (228, 229) and are still widely used for bacterial taxonomic identification with high resolution (230).

These techniques produce massive amounts of data (227) and biostatistical techniques are required to decipher the community response to environmental stressor or time. Strategies consist in exploring diversity through characterising indices (α-diversity), decipher the community structure (β-diversity) and assess the functionality of its members. α-diversity permits the characterisation of the community with the number of species present and their quantitative representation. The richness index measures the number of species. Shannon- Weaver and Simpson indices include the relative abundance of species. ACE and Chao-indexes estimate richness giving more rate to rare species (230). β-diversity positions a local community structure relative to regional landscape.

The impact of environmental disturbance on the community structure can be assessed by focusing on groups of species responding significantly and similarly to the disturbing factor. This ecological concept of “Functional Response Group” (FRG, Box 2) was first conceived for plants (231) and was tried to be linked to “Functional Effect Group” (species with a similar effect on one or several ecosystem function) to explain how the selection of traits can directly or indirectly have functional consequences for the ecosystem (Figure 16). They predict that overlap between traits determining response and effect will be more substantial for biogeochemical trait-filters than stress filters (232). The ecological concept has been used to describe bird (233), ant (234), phytoplankton (235, 236) and bacterial communities. Studies focussing on bacterial communities decipher filtering conditions, such as metals (85), WWTP process (41), drought (237), or fertilisation (238).

Figure 16: Environmental changes impact ecosystem functioning by selecting organisms with specific functional traits changing both the structure of the hosted community and consequently, emergent properties from joint metabolisms of microbial communities. Obtained from Lavorel and Garnier, 2002 (232). Nunes and colleagues (2016) identified 6 functional groups in soil microbial communities responding to copper legacy based on 16S rRNA amplicon sequencing from RNA extracts to assess the active part of the community whose α-diversity indices decrease with metal contamination (85). They found different levels of metal tolerance and used previously referenced bacterial traits to discuss the metal-resistance of the members of those groups (239).

When hypotheses have been drawn from observations in-situ, microcosm or mesocosm monitoring is a tool to confirm those hypotheses and environmental conditions responsible for those observations that could be further tested in synthetic communities. In the Box 2. A Functional Response Group (FRG) is a group of operational taxonomic present work, we consider microcosms units similarly and significantly responding to and mesocosms as simplified, physical a condition. In the present work, significantly impacted bacterial taxa are pinpointed by model of an ecosystem (240). Microbial modelling the microbial communities with a communities are extracted with their negative binomial Generalized Linear Model environmental matrix (i.e., sediments), (nbGLM) and analysing deviance. Chosen OTUs are plotted in a heatmap and FRG placed into a container and monitored in formed by OTUs similarly behaving according the lab in controlled conditions. These to the tested condition, are manually marked micro-ecosystems are called microcosms off. FRG membership of each OTU is finally validated using a Monte-Carlo simulation (85). or mesocosms depending on their size.

They have been widely used to study the effect of contaminants (e.g. metals, pesticide or herbicide), abiotic conditions (e.g. spatial heterogeneity) or stressors (e.g. high temperature) and to study ecological theories, such as competitive exclusion, succession or self-organisation (240). To monitor microcosms, the same ‘omics’ and gene targeting techniques can be applied to assess the microcosm evolution through time.

6.3. In-vitro conjugation assay to study plasmid permissiveness of a natural microbial community The following text was taken from the article: Pinilla-Redondo, R.; Cyriaque, V.; Jacquiod, S.; Sørensen, S.J.; Riber, L. Monitoring Plasmid-Mediated Horizontal Gene Transfer in Microbiomes: Recent Advances and Future Perspectives. Plasmid 2018, 99, 56–67. To overcome the intrinsic weaknesses of the cultivation-dependent methods [for plasmid conjugation event detection], a reporter-gene technology was developed (Figure 17, Box 3). This approach allowed the detection and estimation of plasmid harbouring cells in-situ (both cultivable and non-cultivable), exclusively through the expression of plasmid-encoded fluorescent reporter-genes. Conveniently, these proteins are universally expressed, stable over extended periods of time, and do not require any additional substrates (241). The introduction of fluorescence-based single reporter-gene technologies represented a significant leap forward. These techniques not only allowed the circumvention of the former cultivation biases, but also unlocked the possibility for an upgrade towards higher technologies, like flow cytometry (199, 242). Soon, other improvements followed, like the development of a flow cytometry optimised GFP variant, gfpmut3 (243), which is still extensively used for monitoring the fate of plasmids in natural environments. The primary hallmark of flow cytometry (Box 4) in microbiology resides in its capabilities to enumerate large bacterial populations, at the single- cell level, with overwhelming speeds of several thousand cells per second. Additionally, multiple parameters can be analysed simultaneously, such as a particle's relative size, internal complexity (or granularity), and fluorescence intensity. Together, the combination of features makes flow cytometry a highly useful tool, ideal to discriminate rare events, such as is the case for the typically low-frequency transconjugant cell sub-populations. As a result, the advantages posed by this technique greatly simplify the experimental workflow, sample handling times, and the precision of transfer frequency estimates. Curiously, through the use of this technology, plasmid transfer frequencies were found to be orders of magnitude higher than previously calculated, supporting the view that traditional cultivation-dependent techniques likely underrate the frequency of plasmid-mediated HGT in natural environments (199, 244, 245). Through amplification and sequencing of the 16S rRNA gene, it became possible to uncover the

taxonomic identity of the sorted transconjugant cells (Figure 17C) (199, 245, 246), hence providing an un-precedented picture of the indigenous members of bacterial communities able to receive plasmids. One of the first studies to apply this methodology aimed to examine the transfer range of a broad host IncP plasmid in a soil microcosm. Here, it was demonstrated that plasmid transfer occurred between phylogenetically distant Gram-positive and Gram-negative recipient bacteria, indigenous to the barley rhizosphere (245). The ability of IncP plasmids to undergo trans-Gram transfer in a soil bacterial community was confirmed in another study using single-cell Fluorescence Activated Cell Sorting (FACS) coupled to whole genome amplification of transconjugants (247). This study also revealed that new bacteria were identified as transconjugants when using cultivation-independent techniques, indicating the potential of this method to elucidate a hitherto undetected diversity of transconjugants. However, the numbers of transconjugant cells obtained in both studies were restricted to only a few hundred cells (245, 247). It is worth mentioning the recent introduction of a constitutively expressed fluorescent reporter-gene (e.g. mCherry or dsRed) in the chromosome of plasmid- donor strains (191, 248), in addition to the chromosomally-repressed plasmid-encoded fluorescent marker (e.g. gfp) (Figure 1A). This dual-labelling fluorescence reporter-gene platform facilitates the detection of plasmid transfer via microscopy, flow cytometry, or FACS, through the simultaneous visualisation and/or rapid analysis of donors (red), recipients (non- green-non-red), and transconjugant (green) cells (Figure 1A and B) (199).

In addition to the characterisation of the isolated transconjugant cells, further analyses have proven useful to gain a deeper insight into potential plasmid propagation routes within and between environmental microbiomes. Such post-sorting approaches are based on the analysis of the DNA/RNA profiles of transconjugant pools of different environments and involve correlation and network analyses of the core-permissive fraction (41). For example, the integration of the Functional Response Groups (FRGs) ecological concept represents a promising approach. This notion has been adjusted for clustering operational taxonomic units (OTUs) obtained from 16S rRNA amplicon profiles that respond similarly to environmental factors. These microbial response groups use DNA and RNA abundance signatures as a proxy for microbial presence and activity, respectively (41, 85, 249). Combining the bacterial mating assays and the response group concept allows for a better understanding of transconjugant ecology by categorising them into functionally relevant groups (e.g. copiotrophs/oligotrophs, metal tolerant/sensitive) within the environment. In addition, the metagenomic signature of the in-situ communities supports assertions by pointing out enriched functions (41, 249).

Metagenomics, as well as metaproteomics, could contribute to validate assumptions deducted from the presence and/or activity of specific strains that were highlighted by the MRGs, thereby strengthening the analysis (250).

Box 3. A Conjugation Mating Assay consists in mating a cultivated plasmid donor with a cultivated recipient strain or a recipient community extracted from the environment (Figure 17A). Defined number of cells are mixed and incubated over a fixed period of time. After incubation, cells are resuspended in PBS buffer and the amount of transconjugants, recipient and donor cells are determined. In all systems used in the present work, conjugated plasmid is carrying a gfpmut3 fluorescent marker to detect transconjugants by flow cytometry (Figure 17B, Box 4). Plasmid donor is either detected by a chromosome-encoded mCherry gene associated with a gfpmut3 repressor (Figure 17A&B) or by the Fluorescence in-situ Hybridization (FISH) technique targeting its 16S rRNA gene. In doing so, the number of transconjugants, recipient and donor cells can be assessed and all transconjugant cells of an environmental community can be sorted for 16S rDNA amplicon sequencing (Figure 17C).

Box 4. The Flow cytometry technique combines fluidics, optical and electronic systems to measure parameters associated to single-cells of a whole population. The fluidic system permits to align cells in a capillary tube. When passing in that tube, cells are individually analysed by an optical system. Lasers focus to the cell either to identify its morphology or to detect emitted fluorescence. Morphology is characterised by (i) its size (forward scatter, FSC), that is proportional to the angle degree formed by the laser beam when it is deviated by the cell and (ii) its refringence determined by particulate elements (e.g. mitochondria, granules) (Side scatter; SSC). This factor is estimated by

Figure 17: Studying plasmid permissiveness involves a consecutive plasmid mating assay between the plasmid donors and recipients (A), FACS analysis sorting transconjugant cells (B) and post-sorting DNA extraction, amplification and sequencing to determine the taxonomic identity of transconjugants (C). Obtained from Pinilla-Redondo; Cyriaque et al.; “Monitoring Plasmid-Mediated Horizontal Gene Transfer in Microbiomes: Recent Advances and Future Perspectives”, Plasmid 2018, 99, 56–67.

6.4. Study bacterial interactions in bacterial communities To dig deeper, synthetic communities may be used to decrease the complexity of the system by using pure cultures of bacteria in defined abiotic conditions to understand a metabolism, test stress resistance, assess factors regulating bacterial diversity and interactions in order to decipher observations obtained in-situ and in microcosm or in an engineering purpose.

For instance, Lopatkin and colleagues (2016) assessed the impact of antibiotic concentration on conjugation dynamics. They showed that plasmid transfer rate was mainly dependent on the physiological state of the cells and energy availability and that antibiotics might have more impact on the settlement of the plasmid in the recipient cell by selection (17). Recently, Nolivos and colleagues used a F-Tn10 plasmid transfer reporter based on the fluorescent parS/parBDNA labelling system to show F-Tn10 plasmid transfer between two strains of Escherichia coli (internalisation of ssDNA, re-circularisation, and conversion into dsDNA) lasts about 10 minutes (251). A tetA-mCherry fusion protein and derivatives encoded on the plasmid also allowed them to monitor the TetA production transcription/translation time- lapse (in the following minutes after dsDNA formation) and to highlight the zygotic induction of plasmid genes. In the absence of tetR encoded by the chromosome, TetA is constitutively produced. Finally, they used E. coli mutants (acrA, acrB, or tolC gene deletion) to show their involvement in Tc detoxification in the meantime of TetA synthesis after plasmid acquisition. As tetracycline induces ribosome stalling, they used proteomics to show the decreased amount of protein formation in mutants (251). That study highlights the potential of monitoring techniques by fluorescence and proteomics in the study of HGT in synthetic consortia.

7. Aim of the study The Deûle river (France), contaminated by metals (mainly Cd, Cu, Pb, Zn) released by the foundry MetalEurop over the last century, was compared to upstream control sediments (Férin). Characterisation of this metal-impacted sediment-associated microbial community highlighted their surprising high diversity after more than 100 years of anthropogenic contamination. The central objective of the present work was then to decipher mechanisms that permitted the resilience of a sediment microbial community facing a metal contamination.

We first analysed the taxonomic profile of the microbial communities of Férin and MetalEurop in-situ. First chapter aimed then at drawing hypotheses to elucidate the adaptive tolerance range of the sediment microbial communities. The addressed questions were the following: were all bacterial members evenly selected by metals? What impact had metals on dominant bacteria? Is the microbial structure preserved in sediments downstream of the metal- contamination? What role potentially enriched bacteria would play in the microbial community and how metal-resistance was acquired by micro-organisms?

Hypotheses drawn from in-situ characterization were then tested in-vitro.

The in-situ characterisation only depicted the microbial community after a century of metal contamination, considering these 100-years as a black box. Therefore, a microcosm study (chapter 2) addressed the microbial successional trajectory of sediment-associated microbial community at the early stage of a metal-stress. Does sensitive and tolerant bacteria disappear and get enriched at the same time-scale? What groups are repressed/favoured by metals? Are cooperation or competition drivers of the community succession?

Conjugative plasmids are vessels of accessory genes contributing to genome novelties, making them important vectors of bacterial evolution by accelerating adaptation. Some were shown to carry metal-resistance genes (MRG), including IncP plasmids that were suggested to be ecological key players in the adaptation of metal-impacted microbial communities. To address the role of broad-host range plasmids in the resilience of the community, we assessed the permissiveness (i.e. ability of a community to receive an exogenous IncP plasmid in terms of transfer frequency and diversity of the transconjugants) of the SMCs from Férin and MetalEurop (Chapter 3). Are sediments hotspot for plasmid dispersion? What fraction of the communities is permissive to an exogenous plasmid? Are transconjugant taxonomically diverse?

Finally, the broad host-range IncP plasmid content was also followed by quantitative PCR in sediments placed in microcosms (chapter 2) and measured in-situ (chapter 3). It revealed differential enrichment of these plasmids depending on the time lapse of metal exposure. Then, in-vitro synthetic communities were used to address the impact of lead (Pb) on plasmid dispersal. Does lead repress or favoured the settlement of an exogenous plasmid in a recipient population? Is this effect host- and metal-pressure-dependent? What impact does lead operate on the conjugative machinery? Are beneficial MRGs required for plasmid maintenance? These questions were addressed in chapter 4.

8. Sites of interest: Férin and MetalEurop river sediment community models Massive industrialisation along river basins means that large amounts of hazardous waste, including metals in aquatic sediments, has been discharged (252, 253). Smelting factories produced dust and slag waste which contain high amount of metals. Many of these metals end up in the environment, either through the atmosphere or when they are transported during barge discharge next to the quays (Figure 18 b & c) (254, 255). Agricultural and homegrown crops, as well as environmental soil and sediments, are thus exposed to metal contamination (253, 256).

The present work focuses on river sediments sampled in the “Escaut” watershed (North of France). The studied sediments are situated in the “Deûle” river in Noyelles-Godault, in the “Nord-Pas de Calais” region (31). The Deûle river is canalised along its course and frequently navigated. It is situated next to the former “MetalEurop Nord” foundry, that was one of the largest lead producers of Europe (Figure 18). Along with the active “Umicore” foundry, MetalEurop affected an area of about 120 km² (257). MetalEurop was closed in 2003 after a century of activity, leaving waste materials on the surrounding lands (50, 252). These surroundings still contain high concentrations of metals (254) especially Cd, Pb, Zn and Cu (50, 252, 255). The geo-accumulation indices (Igeo)3 of each metal in suspended matter were respectively 7.9-8.9; 3.5-5.5; 4.8-4.9;1.8-3.2 (255). The probable effect quotient (PEQ)4 sum of these metals ranges from 21.9 to 147.0 in the water and sediments of the area (Figure 19), with a relative amount of metal pollution of the Deûle river as follows: Cd>Pb>Zn >>Cu (31). According to X-ray diffraction analysis (XDR) analyses, MetalEurop sediments are primary composed of calcite (CaCO3), quartz/sand (SiO2) and clays (kaolinite [Al2Si2O5(OH)5] and montmorillonite [Ca0.2(Al,Mg)2Si4O10(OH)2,4H2O]). Sediments also contain lower amounts of hematite (Fe2O3), pyrite (FeS2), galena (PbS), wurtzite (ZnS) and organic matter. Environmental scanning electron microscopy coupled with energy dispersive X-ray spectrometry (ESEM/EDS) analyses revealed that PbS crystals and ZnS-rich particles are associated with Fe oxides, PbCO3 and Fe-rich aggregates. Sediment carbonates seem to play a key role in metal retention (258). Sulphates are also present in high amount at the surface (259).

3 Base 2 logarithm of the measured total concentration of the metal over its background concentration. 15 = extremely polluted. (260) 4 Total sedimentary concentration divided by the probable effect concentration (PEC: concentration at which adverse effects to aquatic biota are expected) 46

a b

Figure 18: Photographs (from September 2005) of the MetalEurop demolition (a) and quays (b, c). Published on http://perso.orange.fr/culture.industrielle, visited on the 6th May 2019).

Figure 19: Additive Probable Effect Quotient (PEQ) of Cd, Pb and Zn in MetalEurop sediment at different sediment depth obtained from Louriño-Cabana et al., 2011 (31). The PEQ of Cu is not represented, as it is less than 1.3 (31). (a)

Metal Reference soilsa Less contaminated Most contaminated sediment upstream sediments in Loessic Alluvial MetalEurop MetalEurop Soils soils (mg/kg)b (mg/kg)b (mg/kg) (mg/kg) Pb 38,4 45 26,8 10079 Zn 73,7 78 149,1 12895 Cd 0,42 0,63 1,15 1399 Ti 0,25 0,32 0,38 227,1 In 0,035 0,029 0,06 75 Sn 2,75 1,76 6,56 78,4 a: reported by Sterckeman et al. (2002), 261 b: reported by Boughriet et al. (2007), 258

Figure 20: Total Metal concentration in MetalEurop sediments, upstream sediments and (254) and XDR analysis of sediment particles of MetalEurop (258) (b). As a control, MetalEurop sediments were compared to upstream sediments sampled in the Sensée canal in Férin (North of France; Figure 21). This area has similar geochemical properties (50) but was not impacted by the MetalEurop foundry, as 70% of the surrounding soils are dedicated to farming activities. Total metal concentrations observed in MetalEurop sediments are up to 30 times the values observed in Férin (Table 1). The highest metal concentrations concern cadmium (Cd), copper (Cu), lead (Pb) and zinc (Zn) (Table 1). Cadmium, lead and zinc are the most represented in the bioavailable fractions obtained by the

two first steps of metal-extraction from sediments (Figure 22; 253). These steps allow the extraction of metals bound to: (i) the carbonates and easily volatile sulphides and (ii) the remaining volatile sulphides, as well as iron and manganese (oxo)hydroxides. The non- bioavailable fractions in which metals are associated are refractory organic, oxides and pyritic compounds, as well as alumina-silicates likes clays (80). The resulting Toxicity Indices (TIs)5 were 0.17 in Férin sediments and 0.84 in MetalEurop sediments. The Metal Pollution Indices (MPIs)6 were 65-72 in Férin and 261 in MetalEurop sediments (38).

(a)

(b)

Figure 21: Localisation of Férin and MetalEurop sites situated on the Sensée canal (Férin) and on the Deûle river (MetalEurop) in the Escaut watershed in north France. Map (a) was obtained modified from Louriño-Cabana et al., 2011 (31). Satellite view (b) was obtained from Google Maps online software (https://www.google.be/maps). Arrow represents the waterflow direction.

5 Ratio bioavailable metals/ AVS 6 MPI=(Cm1*Cm2*…*Cm n)/n. This index evaluates metal content 49

Table 1: Total metal concentration in Férin and MetalEurop sediments obtained by Gillan et al., 2015 (50).

Férin MetalEurop Al (g/kg) 17,6±3 26,2±0,7 As (mg/kg) 2,8±0,3 21±0,9 Cd (mg/kg) 1,3±0,03 38,1±0,5 Co (mg/kg) 5,8±0,01 8,8±0,3 Cr (mg/kg) 56,2±1,7 107,4±1,9 Cu (mg/kg) 13,7±0,4 100±0,8 Fe (mg/kg) 12±0,4 20,6±0,3 Mn (mg/kg) 293,5±7,6 547,9±1,4 Ni (mg/kg) 15,2±0,7 25,5±0,4 Pb (mg/kg) 111,6±0,8 913,8±11 V (mg/kg) 35,9±0,4 62,1±0,6 Zn (mg/kg) 348,5±6,7 3218,5±69

Figure 22: Cumulative Cu, Cd, Pb and Zn concentrations of Férin (FER) and MetalEurop (MET) dry weight sediments obtained by Roosa et al. (2014) (253). White bar represents bioavailable elements obtained from two first steps of metal extraction (exchangeable carbonates and volatile sulphides). Grey bar represents metal obtained from the two-last step of metal extraction (refractory sulphide minerals and residuals metals) that are not bio-available. Barcharts also include cumulative metal concentrations measured in the marine Station 130 (North Sea) and Râches (RAC) sediments sampled in the metal contaminated Scarpe river.

A first insight into the impacts of metals on the ecological quality of sediments in the Deûle river was previously obtained using sediment-dwelling invertebrates (chironomid larvae). It was shown that significant amounts of metals bioaccumulated in these organisms compared to Férin sediments, despite the fact that survival and growth were not affected (80). Compared to Férin, the biomass of SMCs found in the sediments of the Deûle river next to the MetalEurop foundry was increased in the first centimetre layer and was similar in the 4-6 cm layers. The difference may be explained by the small increase in resources such as organic matter, nitrate, phosphate, sulphate and ammonium, allowing microbial growth in the Deûle river (80). The taxonomic profile of the community assessed by metagenomics was not impacted, with a Shannon Index of 4.5 and 4.6 in Férin and MetalEurop, respectively. The SMCs were dominated by β-Proteobacteria (44.8% and 46.2% for, Férin and MetalEurop respectively), mostly with Burkholderia (± 7.5%), Rubrivivax (± 4.9%), Leptothrix (± 3.1%) and Cupriavidus (± 2.6%). Actinobacteria (mostly Mycobacterium and Streptomyces), γ- Proteobacteria (Pseudomonas, 3.3% in Férin and 3.7% in MetalEurop) and α-Proteobacteria (mostly Methylobacterium) were also found (Figure 23) (50). Still, a significant difference in Pseudomonas representation was highlighted by quantitative PCR (Q-PCR) with an increasing amount of this genus with sediment metal levels (38).

Figure 23: Representation heatmap displaying most abundant OTUs (>1%) in Férin and MetalEurop sediments obtained by shotgun metagenomics by Gillan et al. 2015 (50). Those two stations are compared to the non-metal-contaminated marine station 130 (St130). Functionally, a Q-PCR approach highlighted a positive relationship between metal concentration and the occurrence of the czcA gene coding for a Cd, Zn and Co efflux pump (253). A shotgun metaproteogenomics study revealed that the MetalEurop SMC was enriched with genes involved in “Cell wall and capsule”, “Virulence, disease and defence” functions that may be related to metal-resistance, as well as “Phage, prophages, transposable elements and plasmid” associated functions (Figure 24a) (50). Authors also highlighted an increased amount of protein related to respiration virulence and carbohydrate metabolisms in MetalEurop SMC (Figure 24a) (50). This work showed that metal resistance genes, such as czc and cus genes, were more abundant in MetalEurop sediments (Figure 24b). The authors propose that adaptative key genes may have been acquired through MGEs, such as plasmids or viruses (50).

Figure 24: Functional metaproteogenomics analysis of MetalEurop and Férin sediments (50). (a) Metagenomics relative difference between Férin (blue) and MetalEurop (grey) of different SEED classifications. (b) Metaproteomics relative difference between Férin (blue) and MetalEurop (grey) of different SEED classifications (p-value >0.05). (c) Comparison between SEED database classes between Férin (blue) and MetalEurop (grey) sediments highlighting sequences homologous to metal resistance systems (red dots). Numbers (1-6) show the position of metal-resistance genes with the difference between proportions.

9. References 1. Doolittle WF. 2017. Darwinizing Gaia. J Theor Biol 0:1–9. 2. Prosser JI, Bohannan BJMBJM, Curtis TP, Ellis RJ, Firestone MK, Freckleton RP, Green JL, Green LE, Killham K, Lennon JJ, Osborn a M, Solan M, van der Gast CJ, Young JPW. 2007. The role of ecological theory in microbial ecology. Nat Rev Microbiol 5:384–392. 3. Van DerHeijden MGA, Hartmann M. 2016. Networking in the Plant Microbiome. PLoS Biol 14 (2): e1002378. 4. Cray JA, Bell ANW, Bhaganna P, Mswaka AY, Timson DJ, Hallsworth JE. 2013. The biology of habitat dominance; can microbes behave as weeds? Microb Biotechnol 6:453–492. 5. Martiny JBH, Bohannan BJM, Brown JH, Colwell RK, Fuhrman JA, Green JL, Horner-Devine MC, Kane M, Krumins JA, Kuske CR, Morin PJ, Naeem S, Øvreås L, Reysenbach AL, Smith VH, Staley JT. 2006. Microbial biogeography: Putting microorganisms on the map. Nat Rev Microbiol 4:102–112. 6. Konopka A. 2009. What is microbial community ecology? ISME J 3:1223–30. 7. Gleason HA. 1926. The Individualistic Concept of the Plant Association. Bull Torrey Bot Club 53:92:210. 8. Alexandrino M, Costa R, Canário AVM, Costa MC. 2014. Clostridia initiate heavy metal bioremoval in mixed sulfidogenic cultures. Environ Sci Technol 48:3378–3385. 9. Ford T, Ryan D. 1995. Toxic metals in aquatic ecosystems: A microbiological perspective. Environ Health Perspect 103:25–28. 10. Gullberg E, Albrecht LM, Karlsson C, Gullberg E, Albrecht LM, Karlsson C, Sandegren L, Andersson DI. 2014. Antibiotics and heavy metals. MBio 5:19–23. 11. Yi Y, Yang Z, Zhang S. 2011. Ecological risk assessment of heavy metals in sediment and human health risk assessment of heavy metals in fishes in the middle and lower reaches of the Yangtze River basin. Environ Pollut 159:2575–2585. 12. Rajendran P, Muthukrishnan J, Gunasekaran P. 2003. Microbes in heavy metal remediation. Indian J Exp Biol 41:935–944. 13. Baker-Austin C, Wright MS, Stepanauskas R, McArthur J V. 2006. Co-selection of antibiotic and metal resistance. Trends Microbiol 14:176–182. 14. Oves M, Mabood Hussain F. 2016. Antibiotics and Heavy Metal Resistance Emergence in Water Borne Bacteria. J Investig Genomics 3:3–5. 15. Jutkina J, Rutgersson C, Flach CF, Joakim Larsson DG. 2016. An assay for determining minimal concentrations of antibiotics that drive horizontal transfer of resistance. Sci Total Environ 548–549:131–138. 16. Klümper U, Dechesne A, Riber L, Brandt KK, Gülay A, Sørensen SJ, Smets BF. 2017. Metal stressors consistently modulate bacterial conjugal plasmid uptake potential in a phylogenetically conserved manner. ISME J 11:152–165. 17. Lopatkin AJ, Huang S, Smith RP, Srimani JK, Sysoeva TA, Bewick S, Karig DK, You L. 2016. Antibiotics as a selective driver for conjugation dynamics. Nat Microbiol 1:1–8. 18. Walker B, Holling CS, Kinzig A. 2005. Resilience, Adaptability and Transformability in Social– ecological Systems. Phys Rev Lett 9 (2):5. 19. Battin TJ, Besemer K, Bengtsson MM, Romani AM, Packmann AI. 2016. The ecology and biogeochemistry of stream biofilms. Nat Rev Microbiol 14:251–263.

20. Brils J. 2004. The SedNet Strategy Paper The opinion of SedNet on environmentally , socially and economically viable sediment management. Eur Sediment Res Netw 1–13. 21. Apitz SE. 2012. Conceptualizing the role of sediment in sustaining ecosystem services: Sediment-ecosystem regional assessment (SEcoRA). Sci Total Environ 415:9–30. 22. Wohl E, Bledsoe BP, Jacobson RB, Poff NL, Rathburn SL, Walters DM, Wilcox AC. 2015. The natural sediment regime in rivers: Broadening the foundation for ecosystem management. Bioscience 65:358–371. 23. Vanote RL, Minshall WG, Cummins K enneth W, Sedell JR, Cushing CE. 1980. The River Continuum Concept. Can J Fish Aquat Sci 37:130–137. 24. Lyautey E, Jackson CR, Cayrou J, Rols JL, Garabétian F. 2005. Bacterial community succession in natural river biofilm assemblages. Microb Ecol 50:589–601. 25. Mansour I, Heppell CM, Ryo M, Rillig MC. 2018. Application of the Microbial Community Coalescence Concept to Riverine Networks. Biol Rev. . 93 (4): 1832-1845 26. Altermatt F. 2013. Diversity in riverine metacommunities: A network perspective. Aquat Ecol 47:365–377. 27. Bahar MM, Ohmori H, Yamamuro M. 2008. Relationship between river water quality and land use in a small river basin running through the urbanizing area of Central Japan. Limnology 9:19–26. 28. Ou D, Chen B, Bai R, Song P, Lin H. 2015. Contamination of sulfonamide antibiotics and sulfamethazine-resistant bacteria in the downstream and estuarine areas of Jiulong River in Southeast China. Environ Sci Pollut Res 22:12104–12113. 29. Sakan S, Ostojić B, Đorđević D. 2017. Persistent organic pollutants (POPs) in sediments from river and artificial lakes in Serbia. J Geochemical Explor 180:91–100. 30. Lesven L, Lourino-Cabana B, Billon G, Recourt P, Ouddane B, Mikkelsen O, Boughriet A. 2010. On metal diagenesis in contaminated sediments of the Deûle river (northern France). Appl Geochemistry 25:1361–1373. 31. Louriño-Cabana B, Lesven L, Charriau A, Billon G, Ouddane B, Boughriet A. 2011. Potential risks of metal toxicity in contaminated sediments of Deûle river in Northern France. J Hazard Mater 186:2129–2137. 32. Charriau A, Lesven L, Gao Y, Leermakers M, Baeyens W, Ouddane B, Billon G. 2011. Trace metal behaviour in riverine sediments: Role of organic matter and sulfides. Appl Geochemistry 26:80–90. 33. Marti E, Jofre J, Balcazar JL. 2013. Prevalence of Antibiotic Resistance Genes and Bacterial Community Composition in a River Influenced by a Wastewater Treatment Plant. PLoS One 8:1–8. 34. Naura M, Hornby DD, Collins AL, Sear DA, Hill C, Jones JI, Naden PS. 2016. Mapping the combined risk of agricultural fine sediment input and accumulation for riverine ecosystems across England and Wales. Ecol Indic 70:209–221. 35. Sherameti I. 2015. Heavy Metal Contamination of Soils. Springer International Publishing. 36. Proia L, Lupini G, Osorio V, Pérez S, Barceló D, Schwartz T, Amalfitano S, Fazi S, Romaní AM, Sabater S. 2013. Response of biofilm bacterial communities to antibiotic pollutants in a Mediterranean river. Chemosphere 92:1126–1135. 37. Gibbons SM, Jones E, Bearquiver A, Blackwolf F, Roundstone W, Scott N, Hooker J, Madsen R, Coleman ML, Gilbert JA. 2014. Human and environmental impacts on river sediment

microbial communities. PLoS One 9:1–9. 38. Roosa S, Wauven C Vander, Billon G, Matthijs S, Wattiez R, Gillan DC. 2014. The Pseudomonas community in metal-contaminated sediments as revealed by quantitative PCR: A link with metal bioavailability. Res Microbiol 165:647–656. 39. Atashgahi S, Aydin R, Dimitrov MR, Sipkema D, Hamonts K, Lahti L, Maphosa F, Kruse T, Saccenti E, Springael D, Dejonghe W, Smidt H. 2015. Impact of a wastewater treatment plant on microbial community composition and function in a hyporheic zone of a eutrophic river. Sci Rep 5:1–13. 40. Hanna DEL, Tomscha SA, Ouellet Dallaire C, Bennett EM. 2017. A review of riverine ecosystem service quantification: Research gaps and recommendations. J Appl Ecol 1–13. 41. Jacquiod S, Brejnrod A, Morberg SM, Abu Al-Soud W, Sørensen SJ, Riber L. 2017. Deciphering conjugative plasmid permissiveness in wastewater microbiomes. Mol Ecol 26:3556–3571. 42. Brown BL, Swan CM, Auerbach DA, Campbell Grant EH, Hitt NP, Maloney KO, Patrick C. 2011. Metacommunity theory as a multispecies, multiscale framework for studying the influence of river network structure on riverine communities and ecosystems. J North Am Benthol Soc 30:310–327. 43. Rillig MC, Mansour I. 2017. Microbial Ecology: Community Coalescence Stirs Things Up. Curr Biol 27:R1280–R1282. 44. Rillig MC, Antonovics J, Caruso T, Lehmann A, Powell JR, Veresoglou SD, Verbruggen E. 2015. Interchange of entire communities: Microbial community coalescence. Trends Ecol Evol 30:470–476. 45. De Oliveira LFV, Margis R. 2015. The source of the river as a nursery for microbial diversity. PLoS One 10:1–11. 46. Besemer K. 2015. Biodiversity, community structure and function of biofilms in stream ecosystems. Res Microbiol 166:774–781. 47. Sabadini-Santos E, Da Silva TS, Lopes-Rosa TD, Mendonça-Filho JG, Santelli RE, Crapez MAC. 2014. Microbial activities and bioavailable concentrations of Cu, Zn, and Pb in Sediments from a tropic and eutrothicated bay. Water Air Soil Pollut 225:1949 48. Liu T, Zhang AN, Wang J, Liu S, Jiang X, Dang C, Ma T, Liu S, Chen Q, Xie S, Zhang T, Ni J. 2018. Integrated biogeography of planktonic and sedimentary bacterial communities in the Yangtze River. Microbiome 6:1–14. 49. Staley C, Gould TJ, Wang P, Phillips J, Cotner JB, Sadowsky MJ. 2016. Sediments and Soils Act as Reservoirs for Taxonomic and Functional Bacterial Diversity in the Upper Mississippi River. Microb Ecol 71:814–824. 50. Gillan DC, Roosa S, Kunath B, Billon G, Wattiez R. 2015. The long-term adaptation of bacterial communities in metal-contaminated sediments: A metaproteogenomic study. Environ Microbiol 17:1991–2005. 51. Suriya J, Chandra Shekar M, Nathani NM, Suganya T, Bharathiraja S, Krishnan M. 2017. Assessment of bacterial community composition in response to uranium levels in sediment samples of sacred Cauvery River. Appl Microbiol Biotechnol 101:831–841. 52. Xie Y, Wang J, Wu Y, Ren C, Song C, Yang J, Yu H, Giesy JP, Zhang X. 2016. Using in-situ bacterial communities to monitor contaminants in river sediments. Environ Pollut 212:348– 357. 53. Zhang H, Zheng S, Ding J, Wang O, Liu F. 2017. Spatial variation in bacterial community in

natural wetland-river-sea ecosystems. J Basic Microbiol 57:536–546. 54. Besemer K. 2016. Aquatic Biofilms: Ecology, Water Quality and Wastewater TreatmentAquatic Biofilms: Ecology, Water Quality and Wastewater Treatment. Caister Academic Press, Norfolk, UK. 55. Araya R, Tani K, Takagi T, Yamaguchi N, Nasu M. 2003. Bacterial activity and community composition in stream water and biofilm from an urban river determined by fluorescent in situ hybridization and DGGE analysis. FEMS Microbiol Ecol 43:111–119. 56. Lawrence JR, Chenier MR, Roy R, Beaumier D, Fortin N, Swerhone GDW, Neu T., Greer C. 2004. Microscale and Molecular Assessment of Impacts of Nickel , Nutrients , and Oxygen Level on Structure and Function of River Biofilm Communities Microscale and Molecular Assessment of Impacts of Nickel , Nutrients , and Oxygen Level on Structure and Functio. Appl Environ Microbiol 70:4326–4339. 57. Almstrand R, Daims H, Persson F, Sörensson F, Hermansson M. 2013. New methods for analysis of spatial distribution and coaggregation of microbial populations in complex biofilms. Appl Environ Microbiol 79:5978–5987. 58. Liu W, Røder HL, Madsen JS, Bjarnsholt T, Sørensen SJ, Burmølle M. 2016. Interspecific bacterial interactions are reflected in multispecies biofilm spatial organization. Front Microbiol 7:1–8. 59. Liu W, Russel J, Røder HL, Madsen JS, Burmølle M, Sørensen SJ. 2017. Low-abundant species facilitates specific spatial organization that promotes multispecies biofilm formation. Environ Microbiol 19:2893–2905. 60. Toyofuku M, Inaba T, Kiyokawa T, Obana N, Yawata Y, Nomura N. 2016. Environmental factors that shape biofilm formation. Biosci Biotechnol Biochem 80:7–12. 61. Stoodley P, Sauer K, Davies DG, Costerton JW. 2002. Biofilms as Complex Differentiated Communities. Annu Rev Microbiol 56:187–209. 62. Madsen JS, Burmølle M, Hansen LH, Sørensen SJ. 2012. The interconnection between biofilm formation and horizontal gene transfer. FEMS Immunol Med Microbiol 65:183–195. 63. Dufour D, Leung V, Lévesque CM. 2012. Bacterial biofilm: structure, function, and antimicrobial resistance. Endod Top 22:2–16. 64. Kırmusaoğlu S. 2016. Staphylococcal Biofilms: Pathogenicity, Mechanism and Regulation of Biofilm Formation by Quorum-Sensing System and Antibiotic Resistance Mechanisms of Biofilm-Embedded Microorganisms, p. 189–209. In Microbial Biofilms- Importance and Applications. 65. Mizan MFR, Jahid IK, Ha S Do. 2015. Microbial biofilms in seafood: A food-hygiene challenge. Food Microbiol 49:41–55. 66. Singh R, Paul D, Jain RK. 2006. Biofilms: implications in bioremediation. Trends Microbiol 14:389–397. 67. Wolska KI, Grudniak AM, Rudnicka Z, Markowska K. 2016. Genetic control of bacterial biofilms. J Appl Genet 57:225–238. 68. Jain K, Parida S, Mangwani N, Dash HR, Das S. 2013. Isolation and characterization of biofilm-forming bacteria and associated extracellular polymeric substances from oral cavity. Ann Microbiol 63:1553–1562. 69. Besemer K. 2016. Biodiversity , community structure and function of biofilms in stream ecosystems. Res Microbiol 166:774–781.

70. Haas KL, Franz KJ. 2009. Application of Metal Coordination Chemistry To Explore and Manipulate Cell Biology 4921–4960. 71. Duffus JH. 2002. “Heavy metals” a meaningless term? (IUPAC Technical Report). Pure Appl Chem 74:793–807. 72. Lemire JA, Harrison JJ, Turner RJ. 2013. Antimicrobial activity of metals: Mechanisms, molecular targets and applications. Nat Rev Microbiol 11:371–384. 73. Väänänen K, Leppänen MT, Chen XP, Akkanen J. 2018. Metal bioavailability in ecological risk assessment of freshwater ecosystems: From science to environmental management. Ecotoxicol Environ Saf 147:430–446. 74. Magrisso S, Erel Y, Belkin S. 2008. Microbial reporters of metal bioavailability. Microb Biotechnol 1:320–330. 75. Mulligan CN, Yong RN, Gibbs BF. 2001. Remediation technologies for metal-contaminated soils and groundwater: An evaluation. Eng Geol 60:193–207. 76. Prokop Z, Cupr P, Zlevorova-Zlamalikova V, Komarek J, Dusek L, Holoubek I. 2003. Mobility, bioavailability, and toxic effects of cadmium in soil samples. Environ Res 91:119– 126. 77. International Agency for Research on Cancer :: IARC. 2012. Arsenic, metals, fibres, and dusts. World Health Organization, Lyon. 78. Hettiarachchi GM, Pierzynski GM. 2004. Soil lead bioavailability and in-situ remediation of lead-contaminated soils: A review. Environ Prog 23:78–93. 79. Jarosławiecka A, Piotrowska-Seget Z. 2014. Lead resistance in micro-organisms. Microbiol (United Kingdom) 160:12–25. 80. Roosa S, Prygiel E, Lesven L, Wattiez R, Gillan D, Ferrari BJD, Criquet J, Billon G. 2016. On the bioavailability of trace metals in surface sediments: a combined geochemical and biological approach. Environ Sci Pollut Res 23:10679–10692. 81. Bradl HB. 2005. Sources and Origins of Heavy Metals, p. 1–27. In Bradl, HB (ed.), Heavy Metals in the Environment. Elsevier Ltd. 82. Peng J feng, Song Y hui, Yuan P, Cui X yu, Qiu G lei. 2009. The remediation of heavy metals contaminated sediment. J Hazard Mater 161:633–640. 83. Järup L. 2003. Hazards of heavy metal contamination. Br Med Bull 68:167–182. 84. Jaishankar M, Tseten T, Anbalagan N, Mathew BB, Beeregowda KN. 2014. Toxicity, mechanism and health effects of some heavy metals. Interdiscip Toxicol 7:60–72. 85. Nunes I, Jacquiod S, Brejnrod A, Holm PE, Johansen A, Brandt KK, Priemé A, Sørensen SJ. 2016. Coping with copper: Legacy effect of copper on potential activity of soil bacteria following a century of exposure. FEMS Microbiol Ecol 92:1–12. 86. Hobman JL, Crossman LC. 2015. Bacterial antimicrobial metal ion resistance. J Med Microbiol 64:471–497. 87. Antonious GF, Turley ET, Sikora F, Snyder JC. 2008. Heavy metal mobility in runoff water and absorption by eggplant fruits from sludge treated soil. J Environ Sci Heal - Part B Pestic Food Contam Agric Wastes 43:526–532. 88. Alvarenga P, Mourinha C, Farto M, Santos T, Palma P, Sengo J, Morais MC, Cunha-Queda C. 2015. Sewage sludge, compost and other representative organic wastes as agricultural soil amendments: Benefits versus limiting factors. Waste Manag 40:44–52.

89. Tóth G, Hermann T, Da Silva MR, Montanarella L. 2016. Heavy metals in agricultural soils of the European Union with implications for food safety. Environ Int 88:299–309. 90. Tang W, Shan B, Zhang H, Zhang W, Zhao Y, Ding Y, Rong N, Zhu X. 2014. Heavy metal contamination in the surface sediments of representative limnetic ecosystems in eastern China. Sci Rep 4:1–7. 91. Bruins MR, Kapil S, Oehme FW. 2000. Microbial resistance to metals in the environment. Ecotoxicol Environ Saf 45:198–207. 92. Booth SC, Weljie AM, Turner RJ. 2015. Metabolomics reveals differences of metal toxicity in cultures of Pseudomonas pseudoalcaligenes KF707 grown on different carbon sources. Front Microbiol 6:1–17. 93. Ciriolo MR, Civitareale P, Carrì MT, De Martino A, Galiazzo F, Rotilio G. 1994. Purification and characterization of Ag,Zn-superoxide dismutase from Saccharomyces cerevisiae exposed to silver. J Biol Chem 269:25783–25787. 94. Anjem A, Imlay JA. 2012. Mononuclear iron enzymes are primary targets of hydrogen peroxide stress. J Biol Chem 287:15544–15556. 95. Alhasawi A, Auger C, Appanna VP, Chahma M, Appanna VD. 2014. Zinc toxicity and ATP production in Pseudomonas fluorescens. J Appl Microbiol 117:65–73. 96. Chenier D, Beriault R, Mailloux R, Baquie M, Abramia G, Lemire J, Appanna V. 2008. Involvement of fumarase C and NADH oxidase in metabolic adaptation of Pseudomonas fluorescens cells evoked by aluminum and gallium toxicity. Appl Environ Microbiol 74:3977– 3984. 97. French S, Puddephatt D, Habash M, Glasauer S. 2013. The dynamic nature of bacterial surfaces: Implications for metal-membrane interaction. Crit Rev Microbiol 39:196–217. 98. Dibrov P, Dzioba J, Gosink KK, Hase CC. 2002. Chemiosmotic Mechanism of Antimicrobial Activity of Ag+ in Vibrio cholerae. Antimicrob Agents Chemother 46:2668–2670. 99. Nishioka H. 1975. Mutagenic activities of metal compounds in bacteria. Mutat Res Mutagen Relat Subj 31:185–189. 100. Kumar A, Pandey AK, Singh SS, Shanker R, Dhawan A. 2011. Cellular uptake and mutagenic potential of metal oxide nanoparticles in bacterial cells. Chemosphere 83:1124–1132. 101. Chudobova D, Dostalova S, Ruttkay-Nedecky B, Guran R, Rodrigo MAM, Tmejova K, Krizkova S, Zitka O, Adam V, Kizek R. 2015. The effect of metal ions on Staphylococcus aureus revealed by biochemical and mass spectrometric analyses. Microbiol Res 170:147–156. 102. Tiquia-Arashiro SM. 2018. Lead absorption mechanisms in bacteria as strategies for lead bioremediation. Appl Microbiol Biotechnol 102:5437–5444. 103. Izrael-Zivkovic L, Rikalovi M, Gojgic-Cvijovic G, Kazazic S, Vrvic M, Brceski I, Vladimir B, Loncarevic B, Gopcevic K, Ivanka K. 2018. Cadmium specific proteomic responses of a highly resistant Pseudomonas aeruginosa san ai †. RSC Adv 10549–10560. 104. Gillan DC. 2016. Metal resistance systems in cultivated bacteria: Are they found in complex communities? Curr Opin Biotechnol 38:123–130. 105. Naik MM, Dubey SK. 2011. Lead-enhanced siderophore production and alteration in cell morphology in a Pb-resistant Pseudomonas aeruginosa strain 4EA. Curr Microbiol 62:409– 414. 106. Braud A, Geoffroy V, Hoegy F, Mislin GLA, Schalk IJ. 2010. Presence of the siderophores pyoverdine and pyochelin in the extracellular medium reduces toxic metal accumulation in

Pseudomonas aeruginosa and increases bacterial metal toleranceemi. Environ Microbiol Rep 2:419–425. 107. Guo J, Kang Y, Feng Y. 2017. Bioassessment of heavy metal toxicity and enhancement of heavy metal removal by sulfate-reducing bacteria in the presence of zero valent iron. J Environ Manage 203:278–285. 108. Zhou Q, Chen Y, Yang M, Li W, Deng L. 2013. Enhanced bioremediation of heavy metal from effluent by sulfate-reducing bacteria with copper-iron bimetallic particles support. Bioresour Technol 136:413–417. 109. Dhami NK, Reddy MS, Mukherjee MS. 2013. Biomineralization of calcium carbonates and their engineered applications: A review. Front Microbiol 4:1–13. 110. Macaskie LE, Bonthrone KM, Yong P, Goddard DT. 2000. Enzymically mediated bioprecipitation of uranium by a Citrobacter sp.: A concerted role for exocellular lipopolysaccharide and associated phosphatase in biomineral formation. Microbiology 146:1855–1867. 111. Loftin IR, Franke S, Roberts SA, Weichsel A, Héroux A, Montfort WR, Rensing C, McEvoy MM. 2005. A novel copper-binding fold for the periplasmic copper resistance protein CusF. Biochemistry 44:10533–10540. 112. Chong LX, Ash MR, Maher MJ, Hinds MG, Xiao Z, Wedd AG. 2009. Unprecedented binding cooperativity between cui and cuii in the copper resistance protein copk from cupriavidus metallidurans CH34: implications from structural studies by nmr spectroscopy and x-ray crystallography. J Am Chem Soc 131:3549–3564. 113. Giner-Lamia J, López-Maury L, Florencio FJ. 2015. CopM is a novel copper-binding protein involved in copper resistance in Synechocystis sp. PCC 6803. Microbiologyopen 4:167–185. 114. Sharma J, Shamim K, Dubey SK, Meena RM. 2017. Metallothionein assisted periplasmic lead sequestration as lead sulfite by Providencia vermicola strain SJ2A. Sci Total Environ 579:359– 365. 115. Hynninen A, Touzé T, Pitkänen L, Mengin-Lecreulx D, Virta M. 2009. An efflux transporter PbrA and a phosphatase PbrB cooperate in a lead-resistance mechanism in bacteria. Mol Microbiol 74:384–394. 116. Argüello JM, González-Guerrero M, Raimunda D. 2011. Bacterial Transition Metal P1B- ATPases: Transport mechanism and roles in virulence. Biochemistry 50:9940–9949. 117. Delmar JA, Su CC, Yu EW. 2015. Heavy metal transport by the CusCFBA efflux system. Protein Sci 24:1720–1736. 118. Møller AK, Barkay T, Hansen MA, Norman A, Hansen LH, Soslash;rensen SJ, Boyd ES, Kroer N. 2014. Mercuric reductase genes (merA) and mercury resistance plasmids in High Arctic snow, freshwater and sea-ice brine. FEMS Microbiol Ecol 87:52–63. 119. Qin J, Rosen BP, Zhang Y, Wang G, Franke S, Rensing C. 2006. Arsenic detoxification and evolution of trimethylarsine gas by a microbial arsenite S-adenosylmethionine methyltransferase 103. 120. Lemire J, Mailloux R, Auger C, Whalen D, Appanna VD. 2010. Pseudomonas fluorescens orchestrates a fine metabolic-balancing act to counter aluminium toxicity. Environ Microbiol 12:1384–1390. 121. Koechler S, Farasin J, Cleiss-Arnold J, Arsène-Ploetze F. 2015. Toxic metal resistance in biofilms: Diversity of microbial responses and their evolution. Res Microbiol 166:764–773. 122. Kalita D, Joshi SR. 2017. Study on bioremediation of Lead by exopolysaccharide producing

metallophilic bacterium isolated from extreme habitat. Biotechnol Reports 16:48–57. 123. Gupta P, Diwan B. 2017. Bacterial Exopolysaccharide mediated heavy metal removal: A Review on biosynthesis, mechanism and remediation strategies. Biotechnol Reports 13:58–71. 124. Gambino M, Cappitelli F. 2016. Mini-review: Biofilm responses to oxidative stress. Biofouling 32:167–178. 125. Taghavi S, Lesaulnier C, Monchy S, Wattiez R, Mergeay M, Lelie D. 2009. Lead(II) resistance in Cupriavidus metallidurans CH34: Interplay between plasmid and chromosomally-located functions. Antonie van Leeuwenhoek, Int J Gen Mol Microbiol 96:171–182. 126. Morillo Pérez JA, García-Ribera R, Quesada T, Aguilera M, Ramos-Cormenzana A, Monteoliva-Sánchez M. 2008. Biosorption of heavy metals by the exopolysaccharide produced by Paenibacillus jamilae. World J Microbiol Biotechnol 24:2699–2704. 127. So NW, Rho JY, Lee SY, Hancock IC, Kim JH. 2001. A lead-absorbing protein with superoxide dismutase activity from Streptomyces subrutilus. FEMS Microbiol Lett 194:93–98. 128. Giller KE, Witter E, Mcgrath SP. 1998. Toxicity of heavy metals to microorganisms and microbial processes in agricultural soils: A review. Soil Biol Biochem 30:1389–1414. 129. Naveed M, Moldrup P, Arthur E, Holmstrup M, Nicolaisen M, Tuller M, Herath L, Hamamoto S, Kawamoto K, Komatsu T, Vogel H-J, Wollesen de Jonge L. 2014. Simultaneous Loss of Soil Biodiversity and Functions along a Copper Contamination Gradient: When Soil Goes to Sleep. Soil Sci Soc Am J 78:1239. 130. Ouyang F, Ji M, Zhai H, Dong Z, Ye L. 2016. Dynamics of the diversity and structure of the overall and nitrifying microbial community in activated sludge along gradient copper exposures. Appl Microbiol Biotechnol 100:6881–6892. 131. Gołebiewski M, Deja-Sikora E, Cichosz M, Tretyn A, Wróbel B. 2014. 16S rDNA pyrosequencing analysis of bacterial community in heavy metals polluted soils. Microb Ecol 67:635–647. 132. Kwon MJ, Yang JS, Lee S, Lee G, Ham B, Boyanov MI, Kemner KM, O’Loughlin EJ. 2015. Geochemical characteristics and microbial community composition in toxic metal-rich sediments contaminated with Au-Ag mine tailings. J Hazard Mater 296:147–157. 133. Chen Y, Jiang Y, Huang H, Mou L, Ru J, Zhao J, Xiao S. 2018. Long-term and high- concentration heavy-metal contamination strongly influences the microbiome and functional genes in Yellow River sediments. Sci Total Environ 637–638:1400–1412. 134. Wang YP, Shi JY, Wang H, Lin Q, Chen XC, Chen YX. 2007. The influence of soil heavy metals pollution on soil microbial biomass, enzyme activity, and community composition near a copper smelter. Ecotoxicol Environ Saf 67:75–81. 135. Gillan DC, Danis B, Pernet P, Joly G, Dubois P. 2005. Structure of Sediment-Associated Microbial Communities along a Heavy-Metal Contamination Gradient in the Marine Environment. Appl Environ Microbiol 71:679–690. 136. Xu X, Zhang Z, Hu S, Ruan Z, Jiang J, Chen C, Shen Z. 2017. Response of soil bacterial communities to lead and zinc pollution revealed by Illumina MiSeq sequencing investigation. Environ Sci Pollut Res 24:666–675. 137. Sutcliffe B, Chariton AA, Harford AJ, Hose GC, Greenfield P, Elbourne LDH, Oytam Y, Stephenson S, Midgley DJ, Paulsen IT. 2017. Effects of uranium concentration on microbial community structure and functional potential. Environ Microbiol 19:3323–3341. 138. Berg J, Brandt KK, Al-Soud WA, Holm PE, Hansen LH, Sørensen SJ, Nybroe O. 2012. Selection for Cu-tolerant bacterial communities with altered composition, but unaltered

richness, via long-term cu exposure. Appl Environ Microbiol 78:7438–7446. 139. Ding Z, Wu J, You A, Huang B, Cao C. 2017. Effects of heavy metals on soil microbial community structure and diversity in the rice (Oryza sativa L. subsp. Japonica, Food Crops Institute of Jiangsu Academy of Agricultural Sciences) rhizosphere. Soil Sci Plant Nutr 63:75– 83. 140. Tipayno SC, Truu J, Samaddar S, Truu M, Preem JK, Oopkaup K, Espenberg M, Chatterjee P, Kang Y, Kim K, Sa T. 2018. The bacterial community structure and functional profile in the heavy metal contaminated paddy soils, surrounding a nonferrous smelter in South Korea. Ecol Evol 8:6157–6168. 141. Suriya J, Mootapally &, Shekar C, Nathani NM, Suganya T, Bharathiraja S, Krishnan M. 2017. Assessment of bacterial community composition in response to uranium levels in sediment samples of sacred Cauvery River. Environ Biotechnol 101:831–841. 142. Yin H, Niu J, Ren Y, Cong J, Zhang X, Fan F, Xiao Y, Zhang X, Deng J, Xie M, He Z, Zhou J, Liang Y, Liu X. 2015. An integrated insight into the response of sedimentary microbial communities to heavy metal contamination. Sci Rep 5:1–12. 143. Costa PS, Reis MP, Ávila MP, Leite LR, De Araújo FMG, Salim ACM, Oliveira G, Barbosa F, Chartone-Souza E, Nascimento AMA. 2015. Metagenome of a microbial community inhabiting a metal-rich tropical stream sediment. PLoS One 10:1–21. 144. Reis MP, Dias MF, Costa PS, Ávila MP, Leite LR, de Araújo FMG, Salim ACM, Bucciarelli- Rodriguez M, Oliveira G, Chartone-Souza E, Nascimento AMA. 2016. Metagenomic signatures of a tropical mining-impacted stream reveal complex microbial and metabolic networks. Chemosphere 161:266–273. 145. Ren Y, Niu J, Huang W, Peng D, Xiao Y, Zhang X, Liang Y, Liu X, Yin H. 2016. Comparison of microbial taxonomic and functional shift pattern along contamination gradient. BMC Microbiol 16:1–9. 146. Hartmann S, Skrobankova H, Drozdova J. 2013. Inhibition of activated sludge respiration by heavy metals, p. 231–235. In 2013 International Conference on Environment, Energy, Ecosystems and Development Inhibition. 147. Paulo LM, Stams AJM, Sousa DZ. 2015. Methanogens, sulphate and heavy metals: a complex system. Rev Environ Sci Biotechnol 14:537–553. 148. Mudhoo S, Kumar A. 2013. Effects of heavy metals as stress factors on anaerobic digestion processes and biogas production from biomass. Int J Environ Sci Technol 10:1383–1398. 149. Compte-port S, Borrego CM, Restrepo-ortiz CX, Diego A De, Rodriguez-iruretagoiena A, Gredilla A, Vallejuelo SF De, Galand PE, Kalenitchenko D, Rols J, Auguet J. 2018. Metal contaminations impact archaeal community composition , abundance and function in remote alpine lakes 20:2422–2437. 150. Burkhardt EM, Bischoff S, Akob DM, Büchel G, Küsel K. 2011. Heavy metal tolerance of Fe(III)-reducing microbial communities in contaminated creek bank soils. Appl Environ Microbiol 77:3132–3136. 151. Li X, Kapoor V, Impelliteri C, Chandran K, Domingo JWS. 2016. Measuring nitrification inhibition by metals in wastewater treatment systems: Current state of science and fundamental research needs. Crit Rev Environ Sci Technol 46:249–289. 152. Park S, Ely RL. 2008. Candidate stress genes of Nitrosomonas europaea for monitoring inhibition of nitrification by heavy metals. Appl Environ Microbiol 74:5475–5482. 153. Magalhães C, Costa J, Teixeira C, Bordalo AA. 2007. Impact of Trace Metals on

Denitrification in Estuarine Sediments of the Douro River estuary , Portugal. Mar Chem 107:332–341. 154. Zhang X, Chen Z, Zhou Y, Ma Y, Ma C, Li Y, Liang Y, Jia J. 2019. Impacts of the heavy metals Cu ( II ), Zn ( II ) and Fe ( II ) on an Anammox system treating synthetic wastewater in low ammonia nitrogen and low temperature : Fe ( II ) makes a difference. Sci Total Environ 648:798–804. 155. Arora NK, Khare E, Singh S, Maheshwari DK. 2010. Effect of Al and heavy metals on enzymes of nitrogen metabolism of fast and slow growing rhizobia under explanta conditions. World J Microbiol Biotechnol 811–816. 156. Yu Z, Gunn L, Wall P, Fanning S. 2017. Antimicrobial resistance and its association with tolerance to heavy metals in agriculture production. Food Microbiol 64:23–32. 157. Pal C, Bengtsson-Palme J, Kristiansson E, Larsson DGJ. 2015. Co-occurrence of resistance genes to antibiotics, biocides and metals reveals novel insights into their co-selection potential. BMC Genomics 16:1–14. 158. Aminov RI. 2010. A brief history of the antibiotic era: Lessons learned and challenges for the future. Front Microbiol 1:1–7. 159. Cromwell GL. 2002. Why and How Antibiotics Are Used in Swine Production. Anim Biotechnol 13:7–27. 160. Allen HK, Stanton TB. 2014. Altered Egos: Antibiotic Effects on Food Animal Microbiomes. Annu Rev Microbiol 68:297–315. 161. Bird K, Boopathy R, Nathaniel R, LaFleur G. 2019. Water pollution and observation of acquired antibiotic resistance in Bayou Lafourche, a major drinking water source in Southeast Louisiana, USA. Environ Sci Pollut Res. 162. Riaz L, Mahmood T, Khalid A, Rashid A, Ahmed Siddique MB, Kamal A, Coyne MS. 2018. Fluoroquinolones (FQs) in the environment: A review on their abundance, sorption and toxicity in soil. Chemosphere 191:704–720. 163. Li S, Shi W, You M, Zhang R, Kuang Y, Dang C, Sun W, Zhou Y, Wang W, Ni J. 2019. Antibiotics in water and sediments of Danjiangkou Reservoir, China: Spatiotemporal distribution and indicator screening. Environ Pollut 246:435–442. 164. Vilca FZ, Angeles WG. 2018. Occurrence of Antibiotics Residues in the Marine Environment. Examines Mar Biol Ocean 2:12–14. 165. Rodriguez-mozaz S, Chamorro S, Marti E, Luis J. 2015. Occurrence of antibiotics and antibiotic resistance genes in hospital and urban wastewaters and their impact on the receiving river. Water Res 69:234–242. 166. Perry JA, Wright GD. 2013. The antibiotic resistance “mobilome”: Searching for the link between environment and clinic. Front Microbiol 4:1–7. 167. Icgen B, Yilmaz F. 2014. Co-occurrence of antibiotic and heavy metal resistance in Kizilirmak River isolates. Bull Environ Contam Toxicol 93:735–743. 168. Mao D, Yu S, Rysz M, Luo Y, Yang F, Li F, Hou J, Mu Q, Alvarez PJJ. 2015. Prevalence and proliferation of antibiotic resistance genes in two municipal wastewater treatment plants. Water Res 85:458–466. 169. Martins VV, Zanetti MOB, Pitondo-Silva A, Stehling EG. 2014. Aquatic environments polluted with antibiotics and heavy metals: A human health hazard. Environ Sci Pollut Res 21:5873–5878.

170. Zhao F, Yang L, Chen L, Li S, Sun L. 2018. Co-contamination of antibiotics and metals in peri-urban agricultural soils and source identification. Environ Sci Pollut Res 25:34063–34075. 171. Pal C, Asiani K, Arya S, Rensing C, Stekel DJ, Larsson DGJ, Hobman JL. 2017. Metal Resistance and Its Association With Antibiotic ResistanceAdvances in Microbial Physiology, 1st ed. Elsevier Ltd. 172. Seiler C, Berendonk TU. 2012. Heavy metal driven co-selection of antibiotic resistance in soil and water bodies impacted by agriculture and aquaculture. Front Microbiol 3:1–10. 173. Gullberg E, Albrecht LM, Karlsson C, Gullberg E, Albrecht LM, Karlsson C, Sandegren L, Andersson DI. 2014. Selection of a Multidrug Resistance Plasmid by Sublethal Levels of Antibiotics and Heavy Metals. MBio 5:19–23. 174. Olendzenski L, Gogarten JP, Siefert JL, Lawrence JG, Retchless AC, Bapteste E, Boucher Y, Bahl MI, Hansen LH, Sørensen SJ, Labbate M, Case RJ, Stokes HW, Huang J, Gogarten JP, House CH, Fournier G, Margulis L, Zhaxybayeva O, Cortez D, Delaye L, Lazcano A, Becerra A, Poptsova M, Beiko RG, Ragan MA, Oregaard G, Larios-Sanz M, Travisano M, Smets BF, Lardon L, Noll KM, Thirangoon K, Raymond J, Siefert JL, Fox GE, Riley MA, Lizotte- Waniewski M, Papke RT, Barlow M, Coombs JM, Sobecky PA, Hazen TH, Sobecky PA, Coombs JM, Andersson JO, Alsmark UC, Sicheritz-Ponten T, Foster PG, Hirt RP, Embley TM, Keeling PJ, Mitreva M, Smant G, Helder J. 2009. Horizontal Gene Transfer : Genomes in FluxLife Sciences. Humana Press, New York. 175. Soucy SM, Huang J, Gogarten JP. 2015. Horizontal gene transfer: Building the web of life. Nat Rev Genet 16:472–482. 176. Lederberg J, Tatum EL. 1946. Novel genotypes in mixed cultures of biochemical mutants of bacteria, p. 113–114. In Symposia on Quantitative Biology. Cold Spring Harbor Laboratory Press. 177. Tatum EL, Lederberg J. 1947. Gene Recombination in the Bacterium Escherichia coli. J Bacteriol 53:673–684. 178. Smillie C, Garcillan-Barcia MP, Francia M V., Rocha EPC, de la Cruz F. 2010. Mobility of Plasmids. Microbiol Mol Biol Rev 74:434–452. 179. Lederberg J. 1952. Cell Genetics and Hereditary Symbiosis. Physiol Rev 32:403–430. 180. Hinnebusch J, Tilly K. 1993. Linear plasmids and chromosomes in bacteria. Mol Microbiol 10:917–922. 181. Norman A, Hansen LH, Sørensen SJ. 2009. Conjugative plasmids: Vessels of the communal gene pool. Philos Trans R Soc B Biol Sci 364:2275–2289. 182. Clark DP, Pazdernik NJ. 2013. Plasmids. Mol Biol e473–e478. 183. Gama JA, Zilhão R, Dionisio F. 2018. Impact of plasmid interactions with the chromosome and other plasmids on the spread of antibiotic resistance. Plasmid 99:82–88. 184. de la Cruz F, Frost LS, Meyer RJ, Zechner EL. 2010. Conjugative DNA metabolism in Gram- negative bacteria. FEMS Microbiol Rev 34:18–40. 185. Koraimann G, Wagner MA. 2014. Social behavior and decision making in bacterial conjugation 4:1–7. 186. Low KB. 2001. Conjugation. Brenner’s Encycl GenetSecond Edi. Academic Press. 187. Bañuelos-Vazquez LA, Torres Tejerizo G, Brom S. 2017. Regulation of conjugative transfer of plasmids and integrative conjugative elements. Plasmid 91:82–89.

188. Goessweiner-mohr N, Arends K, Keller W, Grohmann E. 2014. Conjugation in Gram-Positive Bacteria. Microbiol Spectr 2:1–19. 189. Kurenbach B, Bohn C, Prabhu J, Abudukerim M, Szewzyk U, Grohmann E. 2003. Intergeneric transfer of the Enterococcus faecalis plasmid pIP501 to Escherichia coli and Streptomyces lividans and sequence analysis of its tra region 50:86–93. 190. Smalla K, Jechalke S, Top EM. 2015. Plasmid Detection, Characterization, and Ecology. Microbiol Spectr 3:1–21. 191. Klümper U, Riber L, Dechesne A, Sannazzarro A, Hansen LH, Sørensen SJ, Smets BF. 2015. Broad host range plasmids can invade an unexpectedly diverse fraction of a soil bacterial community. ISME J 9:934–945. 192. Harrison E, Brockhurst MA. 2012. Plasmid-mediated horizontal gene transfer is a coevolutionary process. Trends Microbiol 20:262–267. 193. Carroll AC, Wong A. 2018. Plasmid persistence: costs, benefits, and the plasmid paradox. Can J Microbiol 64:293–304. 194. Moat AG, Foster JW, Spector MP. 2003. Bacterial Genetics: Dna Exchange, Recombination, Mutagenesis, and RepairMicrobial PhysiologyWiley-Liss. New York. 195. Novick RP. 1987. Plasmid Incompatibility. Microbiol Rev 51:381–395. 196. Gregory R, Saunders JR, Saunders VA. 2008. Rule-based modelling of conjugative plasmid transfer and incompatibility. BioSystems 91:201–215. 197. San Millan A, MacLean RC. 2017. Fitness Costs of Plasmids: a Limit to Plasmid Transmission. Microbiol Spectr 5:1–12. 198. Wen Y, Behiels E, Devreese B. 2014. Toxin – Antitoxin systems : their role in persistence , biofilm formation , and pathogenicity. Pathog Dis 70:240–249. 199. Sørensen SJ, Bailey M, Hansen LH, Kroer N, Wuertz S. 2005. Studying plasmid horizontal transfer in-situ: A critical review. Nat Rev Microbiol 3:700–710. 200. Klumper U, Dechesne A, Smets BF. 2015. Protocol for Evaluating the Permissiveness of Bacterial Communities Toward Conjugal Plasmids by Quantification and Isolation of Transconjugants, p. 1–29. In Hydrocarbon and Lipid Microbiology Protocols - Springer Protocols Handbooks. 201. Piotrowska-Seget Z, Cycoń M, Kozdrój J. 2005. Metal-tolerant bacteria occurring in heavily polluted soil and mine spoil. Appl Soil Ecol 28:237–246. 202. Bruins MR, Kapil S, Oehme FW. 2003. Characterization of a small plasmid ( pMBCP ) from bovine Pseudomonas pickettii that confers cadmium resistance 54:241–248. 203. Gupta SK, Shin H, Han D, Hur H-G, Unno T. 2018. Metagenomic analysis reveals the prevalence and persistence of antibiotic- and heavy metal-resistance genes in wastewater treatment. J Microbiol 56:408–415. 204. Dunon V, Sniegowski K, Bers K, Lavigne R, Smalla K, Springael D. 2013. High prevalence of IncP-1 plasmids and IS1071 insertion sequences in on-farm biopurification systems and other pesticide-polluted environments. FEMS Microbiol Ecol 86:415–431. 205. Jechalke S, Dealtry S, Smalla K, Heuer H. 2013. Quantification of IncP-1 plasmid prevalence in environmental: Samples. Appl Environ Microbiol 79:1410–1413. 206. Brown CJ, Sen D, Yano H, Bauer ML, Rogers LM, Van der Auwera GA, Top EM. 2013. Diverse broad-host-range plasmids from freshwater carry few accessory genes. Appl Environ

Microbiol 79:7684–7695. 207. Heuer H, Schmitt H, Smalla K. 2011. Antibiotic resistance gene spread due to manure application on agricultural fields. Curr Opin Microbiol 14:236–243. 208. Akiyama T, Asfahl KL, Savin MC. 2010. Broad-Host-Range Plasmids in Treated Wastewater Effluent and Receiving Streams. J Environ Qual 39:2211–2215. 209. Popowska M, Krawczyk-Balska A. 2013. Broad-host-range IncP-1 plasmids and their resistance potential. Front Microbiol 4:1–8. 210. Grohmann E. 2011. Horizontal Gene Transfer Between Bacteria Under Natural Conditions, p. 163–187. In Ahmad, I, Ahmad, F, Pichtel, J (eds.), Microbes and Microbial Technology: Agricultural and Environmental Applications. 211. van Elsas JD, Bailey MJ. 2002. The ecology of transfer of mobile genetic elements. FEMS Microbiol Ecol 42:187–197. 212. Richaume A, Angle JS, Sadowsky MJ. 1989. Influence of Soil Variables on in-situ Plasmid Transfer from Escherichia coli to Rhizobium fredii. Appl Environ Microbiol 55:1730–1734. 213. Watson SK, Carter PE. 2008. Environmental influences on Acinetobacter sp. strain BD413 transformation in soil. Biol Fertil Soils 45:83–92. 214. Schafer A, Kalinowski J, Simon R, Puhler A. 1990. High-Frequency Conjugal Plasmid Transfer from Gram-Negative Escherichia coli to Various Gram-Positive Coryneform Bacteria. J Bacteriol 172:1663–1666. 215. Wang Q, Mao D, Mu Q, Luo Y. 2015. Enhanced horizontal transfer of antibiotic resistance genes in freshwater microcosms induced by an ionic liquid. PLoS One 10:1–14. 216. Arango Pinedo C, Smets BF. 2005. Conjugal TOL transfer from Pseudomonas putida to Pseudomonas aeruginosa: effects of restriction proficiency, toxicant exposure, cell density ratios, and conjugation detection method on observed transfer efficiencies. Appl Environ Microbiol 71:51–57. 217. Slager J, Kjos M, Attaiech L, Veening JW. 2014. Antibiotic-induced replication stress triggers bacterial competence by increasing gene dosage near the origin. Cell 157:395–406. 218. Top EM, Rore H De, Collard J, Gellens V, Slobodkina G, Verstraete W, Mergeay M. 1995. Retromobilization of heavy metal resistance genes in unpolluted and heavy metal polluted soil. FEMS Microbiol Ecol 18:191-203 219. Musovic S, Klümper U, Dechesne A, Magid J, Smets BF. 2014. Long-term manure exposure increases soil bacterial community potential for plasmid uptake. Environ Microbiol Rep 6:125– 130. 220. Li L, Dechesne A, He Z, Madsen JS, Nesme J, Sorensen SJ, Smets BF. 2018. Estimating the Transfer Range of Plasmids Encoding Antimicrobial Resistance in a Wastewater Treatment Plant Microbial Community. Environ Sci Technol Lett acs.estlett.8b00105. 221. Torsvik V, Øvreås L. 2002. Microbial diversity and function in soil : from genes to ecosystems. Curr Opin Microbiol 2002, 5:240–245. 222. Bengtsson-Palme J, Larsson DGJ. 2016. Concentrations of antibiotics predicted to select for resistant bacteria: Proposed limits for environmental regulation. Environ Int 86:140–149. 223. Sheth RU, Cabral V, Chen SP, Wang HH. 2016. Manipulating Bacterial Communities by in- situ Microbiome Engineering. Trends Genet 32:189–200. 224. Nostrand JD Van, Wu L, Wu W, Huang Z, Gentry TJ, Deng Y, Carley J, Carroll S, He Z, Gu

B, Luo J, Criddle CS, Watson DB, Jardine PM, Marsh TL, Tiedje JM, Hazen TC, Zhou J. 2011. Dynamics of Microbial Community Composition and Function during in-situ Bioremediation of a Uranium-Contaminated Aquifer ᰔ ‡. Appl Environ Microbiol 77:3860– 3869. 225. Xu M, Wu W, Wu L, He Z, Nostrand JD Van, Deng Y, Luo J, Carley J, Ginder-vogel M, Gentry TJ, Gu B, Watson D, Jardine PM, Marsh TL, Tiedje JM, Hazen T, Criddle CS, Zhou J. 2010. Responses of microbial community functional structures to pilot-scale uranium in-situ bioremediation. ISME J 4:1060–1070. 226. Ponomarova O, Patil KR. 2015. Metabolic interactions in microbial communities: Untangling the Gordian knot. Curr Opin Microbiol 27:37–44. 227. Morales SE, Holben WE. 2011. Linking bacterial identities and ecosystem processes: can‘omic’ analyses be more than the sum of their parts? FEMS Microbiol Ecol 75:2–16. 228. Poretsky R, Rodriguez-r LM, Luo C, Tsementzi D, Konstantinidis KT. 2014. Strengths and Limitations of 16S rRNA Gene Amplicon Sequencing in Revealing Temporal Microbial Community Dynamics 9. 229. Shah N, Tang H, Doak TG, Ye Y. 2011. Comparing bacterial communities inferred from 16S rRNA gene sequencing and shotgun metagenomics. Pac Symp Biocomput 165–176. 230. Kim B, Shin J, Guevarra RB, Lee JH, Kim DW, Seol K, Lee J, Kim HB, Isaacson RE. 2017. Deciphering Diversity Indices for a Better Understanding of Microbial Communities. J Microbiol Biotechnol 27:2089–2093. 231. Lavorel S, McIntyre S, Landsberg J, T.D.A F. 1997. Plant functional classifications: from general groups to specific groups based on response to disturbance. Tree 12. 232. Lavorel S, Garnier E. 2002. Predicting changes in community composition and ecosystem functioning from plant traits: revisting the Holy Grail. Funct Ecol 16:545–556. 233. Dehling DM, Fritz SA, Töpfer T, Päckert M, Estler P, Böhning-gaese K, Schleuning M. 2014. Functional and phylogenetic diversity and assemblage structure of frugivorous birds along an elevational gradient in the tropical Andes IBS special issue. Ecography (Cop) 37:1047–1055. 234. Hoffmann BD, Andersen AN. 2003. Responses of ants to disturbance in Australia , with particular reference to functional groups. Austral Ecol 444–464. 235. Alves-de-souza C, Menezes M, Huszar V. 2006. Phytoplankton composition and functional groups in a tropical humic coastal lagoon , Brazil. 236. Qu Y, Wu N, Guse B, Sun X, Fohrer N. 2019. Riverine phytoplankton functional groups response to multiple stressors variously depending on hydrological periods. Ecol Indic 101:41– 49. 237. Meisner A, Jacquiod S, Snoek BL, Hooven FC, Freedman ZB. 2018. Drought Legacy Effects on the Composition of Soil Fungal and Prokaryote Communities 9:1–12. 238. Wolters B, Jacquiod SJA, Sørensen SJ, Widyasari-mehta A, Bech B, Kreuzig R, Smalla K. 2018. Bulk soil and maize rhizosphere resistance genes , mobile genetic elements and microbial communities are differently impacted by organic and inorganic fertilization. FEMS Microbiol Ecol 94 (4): fiy027 239. Martiny AC, Treseder K, Pusch G. 2013. Phylogenetic conservatism of functional traits in microorganisms. ISME J 7:830–838. 240. Matheson FE. 2008. Microcosms. Encycl Ecol. Academic Press. 241. Andersen JB, Sternberg C, Poulsen LK, Bjørn SP, Givskov M, Molin S. 1998. New Unstable

Variants of Green Fluorescent Protein for Studies of Transient Gene Expression in Bacteria New Unstable Variants of Green Fluorescent Protein for Studies of Transient Gene Expression in Bacteria. Appl Environ Microbiol 64:2240–2246. 242. Sørensen S ø J, Sørensen AH, Hansen LH, Oregaard G, Veal D. 2003. Direct detection and quantification of horizontal gene transfer by using flow cytometry and gfp as a reporter gene. Curr Microbiol 47:129–133. 243. Cormack BP, Valdivia RH, Falkow S. 1996. FACS-optimized mutants of the green fluorescent protein (GFP). Gene 173:33–38. 244. Hausner M, Wuertz S. 1999. High Rates of Conjugation in Bacterial Biofilms as Determined by Quantitative in-situ Analysis High Rates of Conjugation in Bacterial Biofilms as Determined by Quantitative in-situ Analysis. Appl Environ Microbiol 65:3710–3713. 245. Musovic S, Oregaard G, Kroer N, Sørensen SJ. 2006. Cultivation-independent examination of horizontal transfer and host range of an IncP-1 plasmid among gram-positive and gram- negative bacteria indigenous to the barley rhizosphere. Appl Environ Microbiol 72:6687–6692. 246. de la Cruz-Perera CI, Ren D, Blanchet M, Dendooven L, Marsch R, Sørensen SJ, Burmølle M. 2013. The ability of soil bacteria to receive the conjugative IncP1 plasmid, pKJK10, is different in a mixed community compared to single strains. FEMS Microbiol Lett 338:95–100. 247. Shintani M, Matsui K, Inoue J ichi, Hosoyama A, Ohji S, Yamazoe A, Nojiri H, Kimbara K, Ohkuma M. 2014. Single-cell analyses revealed transfer ranges of incP-1, incP-7, and incP-9 plasmids in a soil bacterial community. Appl Environ Microbiol 80:138–145. 248. Musovic S, Dechesne A, Sørensen J, Smets BF. 2010. Novel assay to assess permissiveness of a soil microbial community toward receipt of mobile genetic elements. Appl Environ Microbiol 76:4813–4818. 249. Jacquiod S, Cyriaque V, Riber L, Al-soud WA, Gillan DC, Wattiez R, Sørensen SJ. 2018. Long-term industrial metal contamination unexpectedly shaped diversity and activity response of sediment microbiome. J Hazard Mater 344:299–307. 250. Ranjan R, Rani A, Metwally A, McGee HS, Perkins DL. 2016. Analysis of the microbiome: Advantages of whole genome shotgun versus 16S amplicon sequencing. Biochem Biophys Res Commun 469:967–977. 251. Nolivos S, Cayron J, Dedieu A, Page A, Delolme F, Lesterlin C. 2019. Role of AcrAB-TolC multidrug efflux pump in drug-resistance acquisition by plasmid transfer. Science (80- ) 364:778–782. 252. Vdović N, Billon G, Gabelle C, Potdevin JL. 2006. Remobilization of metals from slag and polluted sediments (Case Study: The canal of the Deûle River, northern France). Environ Pollut 141:359–369. 253. Roosa S, Wattiez R, Prygiel E, Lesven L, Billon G, Gillan DC. 2014. Bacterial metal resistance genes and metal bioavailability in contaminated sediments. Environ Pollut 189:143–151. 254. Boughriet A, Proix N, Billon G, Recourt P, Ouddane B. 2007. Environmental impacts of heavy metal discharges from a smelter in Deûle-canal sediments (northern France): Concentration levels and chemical fractionation. Water Air Soil Pollut 180:83–95. 255. Lourino-Cabana B, Lesven L, Billon G, Denis L, Ouddane B, Boughriet A. 2012. Benthic exchange of sedimentary metals (Cd, Cu, Fe, Mn, Ni and Zn) in the Deûle River (Northern France). Environ Chem 9:485–494. 256. Douay F, Pelfrêne A, Planque J, Fourrier H, Richard A, Roussel H, Girondelot B. 2013. Assessment of potential health risk for inhabitants living near a former lead smelter. Part 1:

Metal concentrations in soils, agricultural crops, and homegrown vegetables. Environ Monit Assess 185:3665–3680. 257. Douay F, Pruvot C, Roussel H, Ciesielski H, Fourrier H, Proix N, Waterlot C. 2008. Contamination of urban soils in an area of Northern France polluted by dust emissions of two smelters. Water Air Soil Pollut 188:247–260. 258. Boughriet A, Recourt P, Proix N, Billon G, Leermakers M, Fischer JC, Ouddane B. 2007. Fractionation of anthropogenic lead and zinc in Deûle River sediments. Environ Chem 4:114– 122. 259. Lourino-Cabana B, Lesven L, Billon G, Denis L, Ouddane B, Boughriet A. 2012. Benthic exchange of sedimentary metals (Cd, Cu, Fe, Mn, Ni and Zn) in the Dele River (Northern France). Environ Chem 9:485–494. 260. Odat S. 2015. Application of Geoaccumulation Index and Enrichment Factors on the Assessment of Heavy Metal Pollution along Irbid/zarqa Highway-Jordan. J Appl Sci 15:1318– 1321. 261. Sterckeman T, Douay F, Proix N, Fourrier H, Perdrix E. 2002. Assesment of the contamination of cultivated soils by eighteen trace elements around smelters in the north of France. Water Air Soil Pollut 135:173–194. 262. Wozniak RAF, Waldor MK. 2009. Integrative and conjugative elements: mosaic mobile genetic elements enabling dynamic lateral gene flow. Notfall und Hausarztmedizin 35:115.

Results

Chapter 1: Long-term industrial metal contamination unexpectedly shaped diversity and activity response of sediment microbiome

Journal of Hazardous Materials, 344 (2018) 299-307

Journal of Hazardous Materials 344 (2018) 299–307

Contents lists available at ScienceDirect

Journal of Hazardous Materials

j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / j h a z m a t

Research paper

Long-term industrial metal contamination unexpectedly shaped diversity and activity response of sediment microbiome

Samuel Jacquiod a,∗,1 , Valentine Cyriaque b,1 , Leise Riber c , Waleed Abu Al-soud a , David b b a C. Gillan , Ruddy Wattiez , Søren J. Sørensen a Section of Microbiology, Department of Biology, University of Copenhagen, Universitetsparken 15, 2100 Copenhagen Ø, 1, Bygning, 1-1-215, Denmark b Proteomics and Microbiology Lab, Research Institute for Biosciences, UMONS, avenue du Champs de Mars 6, 7000 Mons, Belgium c Section of Functional Genomics, Department of Biology, University of Copenhagen, Ole Maaløesvej 5, 2200 Copenhagen N, Denmark

highlights

• Combined DNA/RNA sequencing and FRGs accurately predicted microbial lifestyles. • Metal pollution in sediment resulted in unexpected higher microbial diversity.

• Community coalescence, HGT and microbial facilitation explained this higher diversity.

article info abstract

Article history: Received 26 April 2017 Metal contamination poses serious biotoxicity and bioaccumulation issues, affecting both abiotic conditions and biological Received in revised form 11 August 2017 activity in ecosystem trophic levels, especially sediments. The MetalEurop foundry released metals directly into the French Accepted 25 September 2017 Available online river “la Deûle” during a century, contaminating sediments with a 30-fold increase compared to upstream unpolluted areas 28 September 2017 (Férin, Sensée canal). Previous metaproteogenomic work revealed phylogenetically analogous, but functionally different microbial communities between the two locations. However, their potential activity status in-situ remains unknown. The present Keywords: study respectively compares the structures of both total and active fractions of sediment prokaryotic microbiomes by coupling Metals DNA and RNA-based sequencing approaches at the polluted MetalEurop site and its upstream control. We applied the Anthropogenic pollution innovative ecological concept of Functional Response Groups (FRGs) to decipher the adaptive tolerance range of the River sediment communities through characterization of microbial lifestyles and strategists. The complementing use of DNA and RNA Functional response group sequencing revealed indications that metals selected for mechanisms such as microbial facilitation via “public-good” providing 16S rRNA sequencing bacteria, Horizontal Gene Transfer (HGT) and community coalescence, overall resulting in an unexpected higher microbial diversity at the polluted site.

1.Introduction

∗ Corresponding author. Present address: Agroécologie UMR1347, INRA Dijon Centre, 17 rue Sully, Although known to naturally occur due to particular geological context, metal 21000, Dijon, France. contamination of soils and sediments often origi- E-mail addresses: [email protected] (S. Jacquiod), [email protected] (V. Cyriaque), [email protected] (L. Riber), [email protected] (W.A. Al-soud), [email protected] (D.C. Gillan), [email protected] (R. Wattiez), [email protected] (S.J. Sørensen).

1 Samuel Jacquiod and Valentine Cyriaque have contributed equally to this work as shared

300 S. Jacquiod et al. / Journal of Hazardous Materials 344 (2018) 299–307 nates from anthropogenic activities such as mining activities [1,2], wood In the present study, we have revisited the sediments from the Deûle river sites processing [3], shipping, dredging [4], urbanization [5] and industrial processes in northern France, previously investigated by a metaproteogenomic approach [6,7]. These metals constitute a serious risk because of their biotoxicity and [7]. Sediments were exposed to long-term metal releases from the industrial site bioaccumulation in the environment [8]. of MetalEurop, a former foundry operating from 1893 to 2003 in Noyelles- Godault [19]. Metal concentrations in these sediments are currently up to 30- The peculiar characteristics of sedimentary environments, such as redox fold higher compared to control upstream locations at the Sensée canal in Férin. potential, pH, organic matter as well as biological activity, turn them into Sediments are mainly contaminated with cadmium, copper, lead and zinc that natural accumulation hotspots adsorbing and precipitating up to 90% of soluble respectively reach 38.1, 100, 913.8 and 3218.5 mg/kg (Table S1) [6,7,19]. metals and metalloids from water compartments [2,9]. Fresh water sediments Gillan et al. showed by shotgun metagenomics that microbiomes from Férin are hosting thousands of microbial species [10], seating at the bottom of the (FER) and MetalEurop (MET) were phylogenetically analogous but function- trophic levels and often actively involved in metal movements in the biosphere ally different [7]. In this study, we aimed to investigate the sediment prokaryotic [11]. Therefore, many studies have analyzed the link between metal communities with a refined complementing approach using both DNA and accumulation and microbial communities in sediments. Investigating sediment RNA (cDNA) molecular levels by means of high throughput sequencing of the microbiomes represents a good opportunity to understand resistance/tolerance 16S rRNA gene. We hypothesized that the long-term pollution has impacted adaptation and molecular mechanisms involved, with important applications in the prokaryote diversity and selected for different microbial strategies and the field of bioremediation and biostimulation [12]. Metal-contaminated lifestyles. We applied the ecological concept of functional response groups sediment bacteria highlighted by previous studies were mainly affiliated to (FRGs), which aims to classify the response of microorganisms “as a function Proteobacteria [1,7,13,14], Bacteroidetes [13,14], Firmicutes [12,13] and of” environmental parameters [3,20,21]. FRGs should be clearly differentiated Actinobacteria [5,7]. Beta- and Gammaproteobacteria are the essential from Functional Effect Groups (aka guilds), which are groups of organisms Proteobacterial classes, including respectively microbial members from contributing to the same ecosystem function (e.g. nitrogen cycling or cellulose Burkholderiales [2,7,13,14] and Pseudomonadales/Xanthomonadales [5,7,12– degradation). Defining response groups based on RNA/DNA abundance 14]. patterns in relation to environmental variables and without any phylogenetic a priori is a powerful method in ecology for detecting niches occupied by specific microbial strategists [3,20]. Although communities are sharing similarities Total microbial biomass, activity and phylogeny are the most studied traits as previously reported, the resolution of our analysis allowed identification of when investigating the response of environmental microorganisms to metals. six microbial response groups with specific DNA and RNA molecular For instance, Gillan et al. [7] found no changes in biomass and activity in an signatures linked to metal sensitivity/tolerance after long-term exposure. Our 80-years metal contaminated fjord, but revealed community structure study adds a decisive and innovative contribution to the current knowledge variations via DGGE fingerprints, implying long-term adaptation and regarding microbial adaptation to metals in sediments, with regards to the functional recovery over time [15]. These observations were also reported for contrasting results often reported in the literature. other systems, including river-connected lakes in the Rouyn-Noranda region, Canada [16], as well as long-term copper polluted grasslands, for which the microbial community composition was altered at DNA level with no consequences on species richness [17]. Conversely, Nayar and colleagues used mesocosms to show the negative impact of metals on crucial ecosystem 2.Materials and methods components, such as primary producers (e.g. phytoplankton and autotrophic bacterial activity) [4]. They also pointed out that heterotrophic bacteria seem to 2.1. Sampling, DNA and RNA extraction, cDNA synthesis be less affected by metals as a short-term stress [4]. Conversely, some recent studies have revealed strong negative impact of metals on microbial diversity Sediments were sampled in May 2016 from the Sensée Canal and the Deûle in terms of richness and evenness [1,3,5]. These contrasting observations imply river sediment in Férin (FER) and Noyelles-Godault next to MetalEurop (MET) that other factors are involved in structuring the microbial response to metals. in France, respectively. Three sediment cores were collected at each station and These might include site-specific physicochemical differences (e.g. pH, two samples of 2 g were taken from the upper part of each core, representing a organic matter…), sediment sampling depth and associated metal total of 12 samples (3 cores × 2 samples at FER and MET, respectively, Table ◦ bioavailability (e.g. anoxic gradient, ecosystem temporal dynamics…), the S2). Samples were stored in Life-Guard RNA blocking solution (Mobio) at 4 nature of social interactions between microbiome members (e.g. facilitation, C during transport and −20 ◦ C in the laboratory. For DNA/RNA extraction, 6 exclusion, prior-ity, biofilms, keystone species…) and genetic modalities of × 2 g of sediments per station were washed using the Fortin et al. (2004) metal resistance/tolerance (chromosomes and/or plasmids) [18]. In addition, procedure in order to remove potential PCR inhibitors [22]. From the 2 mL re- differences between molecular markers have been reported. For instance, Berg suspended and washed sediment, 500 µL were used for total DNA extraction et al. found no richness loss after long-term copper pollution in grassland soil (FastDNA® SPIN Kit for Soil, MP Biomedicals, Santa Ana, CA, USA) and 800 at the DNA level [17], while a recent study on the same site reported a clear TM µL were used for total RNA extraction (FastRNA Pro Soil-Direct Kit, MP diversity loss in the potentially active microbial fraction at the RNA level [3], Biomedicals, Santa Ana, CA, USA). DNA was removed from the RNA solution both based on 16S rRNA amplicon sequencing. This suggests that some ® TM community members might display low activity profiles, explain-ing the with a DNaseI treatment using the Ambion DNA-free DNase Treatment maintained DNA diversity levels unlike RNA. This also implies that genetic and Removal Reagents kit (ThermoFisher Scien-tific, Waltham, MA, USA). diversity may not be lost due to metal pollution per se, but will still be present cDNA synthesis was performed using 10 ng of DNaseI treated RNA as template with Random Hexamer primers (Sigma, St. Louis, MO, USA) using the Roche under latent state, suggesting potential reactivation of dormant ecosystem TM functions in case of disturbance removal. Overall, these contrasting Expand Reverse Transcriptase kit (Roche, Basel, Switzerland), according to observations call for better alternatives to investigate and understand the manufacturer’s instruction. Generated cDNA samples were stored at −20 ◦ C ecology and modalities of microbial adaptation to metals. until further processing.

S. Jacquiod et al. / Journal of Hazardous Materials 344 (2018) 299–307 301

2.2. High throughput 16S rRNA gene sequencing Table 1 PERMANOVA analysis based on Bray-Curtis dissimilarity with 10, 000 permutations. The top panel shows a three-way PERMANOVA on the following factors: 1) Sediment core, 2) Amplicon sequencing was realized according to acknowledged best practices DNA/RNA level and 3) Site (Férin/MetalEurop). The bottom panel is showing refined one-way guidelines [23]. An initial PCR amplification step was performed, using a set PERMANOVAs showing the nested differences between sites according to 1) DNA and 2) RNA of primers, 341F: 5 -CCTAYGGGRBGCASCAG-3 and 806R: 5 - levels, followed by the differences between RNA/DNA levels within each site, namely 3) Férin GGACTACNNGGGTATCTAAT-3 [3], which flank the approx. 460 bp and 4) MetalEurop. Significance: ***if p < 0.001; **if p < 0.01; *if p < 0.05. variable V3–V4 region of the Prokaryote 16S rRNA gene, including domains of Archaea and Bacteria. Tagging and adding sequencing adapters to amplified Factors tested r2 p Signif. DNA was done in a second amplification step using fusion primers that have Three-way permanova (Bray-Curtis, 10000 perms.) adaptor barcode tags and spacers as previously described [3]. Purification and 1: Sediment core 0.048 0.105 – size-selection (removal of products less than 200 bp) of the approx. 620 bp PCR 2: DNA/RNA level 0.142 1.7E–4 *** amplicon products was performed using Agencourt AMPure XP beads 3: Site (Férin/Metal) 0.224 1.0E–5 *** (Beckman Coulter, Brea, CA, USA) according to manufacturer’s instructions. One-way permanovas (Bray-Curtis, 10000 perms.) The samples were pooled and adjusted to equimolar concentrations, 1: DNA level: Férin vs Metal 0.361 8.2E–3 ** concentrated using the DNA Clean and ConcentratorTM -5 kit (Zymo Research, 2: RNA level: Férin vs Metal 0.321 3.3E–4 *** 3: Férin: DNA vs RNA 0.224 0.015 * Irvine, CA, USA). Finally, they were subjected to 2 × 250 bp paired-end high- 4: Metal: DNA vs RNA 0.291 8.1E–4 *** throughput sequencing on an Illumina® MiSeq® platform (Illumina, San Diego, CA, USA) according to manufacturer’s instructions. Unassembled raw amplicon data were deposited at the Sequence Read Archive public repository the Rgui software version 3.0.2 [36] using the multcomp package with ANOVA (SRA, https://www.ncbi.nlm.nih.gov/sra) under the accession number and a post-hoc Tukey HSD correction test (p < 0.05). SRP112522 (https://www.ncbi.nlm.nih.gov/ Traces/study/?acc=SRP112522). 2.5. Beta-diversity analysis

The dissimilarity and multivariate analysis were done with the Rgui software

[36] using the Rgui package vegan [37] as previously described [3,20,33]. As 2.3. Annotation and generation of the contingency table the OTU contingency tables have features with abundance variation higher than 1000-folds, a log10 trans-formation was applied. A cluster dendrogram based Amplicon analysis was realized according to acknowledged best practices on Bray-Curtis dissimilarity was established on OTU profiles using 1, 000 guidelines [23]. Generated amplicon sequences were analyzed using qiime pipe permutations (Fig. 2). PERMANOVA tests were performed on the Bray-Curtis (https://github.com/maasha/qiime pipe) as previously described [3]. Sequence dissimilarity profiles using 10, 000 permutations (Table 1) to assess the demultiplexing was done using the MiSeq Controller Software and diversity significance of the tested factors (sampling site, nucleotide material, sampling spacers were trimmed using biopieces (www.biopieces.org). Sequence mate- core). Major relative abundance phylogenetic changes in the datasets were pairing and filtering was done using usearch v7.0.1090 . OTU clustering, investigated by means of ANOVA with a false discovery rate correction test dereplication and singleton removal was performed using uparse [25]. (FDR, p < 0.05). Changes attributed to sediment site-specific differences are Paired-end mating was applied with a minimum overlap of 50 bp, maximum presented in Supporting Table S3. Differences related to nucleotide levels mismatches of 15 and a minimum quality of 30. Criteria for sequence trimming (DNA and RNA) are presented in Supporting Table S4. were based on: (1) reads shorter than 200 bp, (2) average quality scores lower than 25, (3) maximum number of ambiguous bases and (4) six as maximum lengths of homopolymers. Chimera checking and removal was performed using 2.6. Identification and validation of functional response groups (FRGs) usearch and the ChimeraSlayer package [26]. Operational Taxonomic Units (OTUs) were picked at 97% sequence identity level using Mothur v.1.25.0 [27]. An UniFrac phylogenetic tree was built using Trainset 9 [28] with QIIME The identification and validation of FRGs was adapted from a previously wrappers for PyNAST [29], Fast-Tree [30], and alignment filtering [31]. A read described procedure [3,20]. OTUs significantly affected between sites were contingency table was exported at species level. Samples with less than 2, 000 identified with an analysis of deviance on the raw non-rarefied counts under sequences were not considered, as they barely provide enough coverage for negative binomial distribution and generalized linear model, corrected by 1, 000 further diversity analysis [32]. Information regarding the sequence counts for resampling iterations of the residual variance (nbGLM, p < 0.05) using the Rgui each sample is provided in the Supporting Table S2. package mvabund [38]. The 139 significantly affected OTUs were plotted in a generalized heatmap, and response groups were defined with a hierarchical cluster dendrogram (Euclidean distance and average clustering) using the Rgui 2.4. Alpha-diversity analysis package vegan (Fig. S4) [37]. Statistical validity of the so-obtained six FRGs was tested against a null-model by Monte-Carlo simulation with all OTUs to Alpha-diversity analysis was carried out as previously described [20,33]. The reinforce randomization power and avoid a priori effects during group raw sequencing counts were used directly to estimate the sequencing depth determination (Fig. S5). A summary showing FRGs abundance patterns and completeness via rarefaction curves (Table S3) using the PAST software [34]. phylogenetic composition is shown in Fig. 3. Diversity indices were calculated on rarefied data at 8, 900 counts per sample. Samples below 8, 900 counts were not included for this particular analysis (Table S2), leaving each condition with at least four biological replicates. Venn Phylogenetic structure in FRGs was tested using relatedness indices [39]. The diagrams were established to display the richness distribution of rarefied data unweighted UniFrac 16S rRNA amplicon tree produced by QIIME was used among the different tested conditions (Fig. S1) using the Rgui package limma with the Rgui package ape [40]. The RGui package picante was used to [35]. The following indices were used to assess the diversity: the sample calculate relatedness [41] via the Mean Nearest Taxon Distance index (MNTD, richness, the Shannon (H), the Chao-1 and ACE indices (Fig. 1). Statistical aka −1*NTI for Nearest Taxon Index) [42]. Significance of the FRGs analysis was done with relatedness indices was tested using a simulated null model with random mock groups of the same size (Z-scores, 10, 000 permutations, 95% confidence interval, p < 0.05, Table 2).

302 S. Jacquiod et al. / Journal of Hazardous Materials 344 (2018) 299–307

Fig. 1. Diversity indices in each sediment sites and according to DNA and RNA levels. Statistical significance was inferred by ANOVA with a Tukey post-hoc correction test (p < 0.05). The letters indicate statistically significant differences between conditions.

Table 2 Phylogenetic relatedness of the Functional Response Groups (FRGs) and overall sediment communities (Férin and MetalEurop) and the associated statistical analysis. The Mean Nearest Taxon Distance (MNTD obs.) was calculated for each FRG and their mock version based on the UniFrac distance in the OTU tree and compared to their respective simulated null model counterpart (MNTD null ±SEM) using a Z-test (Z-scores after a null model simulation with 1, 000 permutations, p < 0.05). Significance: ***if p < 0.001; **if p < 0.01; *if p < 0.05.

FRGs OTUs (n) MNTD (obs.) MNTD (null) MNTD (z-score) MNTD (p)

FRG1 36 0.174 0.236 ± 0.001 −1.841 0.028* FRG2 29 0.245 0.254 ± 0.001 −0.228 0.437 FRG3 39 0.229 0.229 ± 0.001 −0.005 0.496 FRG4 13 0.16 0.335 ± 0.002 −92.303 0.003** FRG5 11 0.275 0.36 ± 0.003 −0.955 0.161 FRG6 11 0.292 0.359 ± 0.003 −0.763 0.243 Férin 314 0.105 0.106 ± 0.003 −0.349 0.353 MetalEurop 360 0.097 0.101 ± 0.002 −2.783 0.003**

3.Results and discussion ecosystem types and site specificities [3,16,17,43], some reported cases for Studies focused on microbial ecology of long-term environmental metal similar locations still revealed interpellant observations depending on applied contaminations have reported contrasting observations in terms of adaptation experimental design, especially the molecular methodology used (e.g. DNA strategies and community diversity status. Although part of these discrepancies [17] or RNA [3]). In this study, the investigated model sediment sites were could be explained by previously studied via metagenomics [7], revealing taxonomically analogous communities

S. Jacquiod et al. / Journal of Hazardous Materials 344 (2018) 299–307 303

Fig. 2. Cluster dendrogram analysis of the taxonomic profiles at OTU level based on the Bray-Curtis dissimilarity index. The grouping coherency was assessed by 1, 000 bootstraps retaining the highest cophenic correlation index possible (0.78).

Fig. 3. Composition and abundance of the six Functional Response Groups identified in this study. Panel A represents the abundance patterns of the six FRGs across the sites (Férin and MetalEurop) and RNA/DNA levels. Panel B represents the weighted abundance of phylogenetic groups in each FRG for each compartment. The grey color in the right barchart represents all remaining sequences not included in the particular FRG displayed. between foundry-contaminated sediments (MetalEurop) and a control upstream level (Fig. S2) and with an overall high beta-diversity similarity level between location. This work goes a step further using 16S rRNA gene sequencing at profiles (Table 1). Indeed, the PERMANOVA on the Bray-Curtis dissimilarity DNA and RNA (cDNA) molecular levels to precisely characterize community index with 10, 000 permutations (Table 1), revealed significant but moderate 2 2 diversity and structure via identification of FRGs for better interpretations of site (r = 0.23, p = 1.0E-5) and DNA/RNA effects (r = 0.14, p = 2.9E-4), 2 previous observations and pinpoint microbial strategies. followed by a minor but significant interaction between the two factors (r = 0.06, p = 0.04). No effect of the sampling core was detected, indicating overruling site and DNA/RNA effects compared to biological sampling 3.1. RNA and DNA description of sediment microbiomes 2 heterogeneity (r = 0.05, p = 0.11). Nested analyses at the nucleotide level Despite significant differences seen on richness between Férin and MetalEurop (DNA vs RNA; Table 1) revealed a quasi-similar and significant discrimi- 2 2 (Fig. 1), taxonomic profiling confirmed previously observed analogies between native power for both DNA (r = 0.36, p = 8.2E-3) and RNA (r = 0.37, p = sites, sharing 70% OTUs at the DNA 2.2E-3) in differentiating the two sites. This is coherent with

304 S. Jacquiod et al. / Journal of Hazardous Materials 344 (2018) 299–307 evenness analysis showing similar Shannon index levels between DNA and mean-to-variance relationship [37]. A total of 139 OTUs responding RNA profiles (Fig. 1). On the other hand, the DNA/RNA differences at each significantly were extracted and represented in a heatmap for pattern detection 2 site seemed to be more explanatory and significant at MetalEurop (r = 0.38, p (Fig. S4). The constrained ordination of all OTUs found in this study under 2 = 2.2E-3) compared to the Férin site (r = 0.22, p = 0.01). This makes sense as Monte-Carlo simulation revealed a very significant and non-random clustering the river dynamic flow between the upstream control site and the downstream into six distinct response groups (called FRG1-6, p < 1.0E-08, Fig. S5) polluted one would continuously contribute in maintaining the microbial displaying peculiar molecular patterns and phylogenetic signatures (Fig. 3, diversity longitudinally with new inoculum material carried by the water stream Table 2). [44]. The dormant “seed bank” (FRG1) represents a ubiquitous fraction of the community, characterized by a significant phylogenetic signal (Table 2) from Proteobacteria was very prominent, which is expected in these ecosystems metal-tolerant but slow-growing and/or inactive members, mainly from [1,7,14]. Gammaproteobacteria (Enterobacteriales, Pseudomonadales and Gamma-, Betaproteobacteria and Bacteroidetes. It features fresh-water bacteria Aeromonadales) clearly dominated all samples, (Fig. 2) although they associated to metal-tolerance, like Aeromonas and Shewanella [55], significantly dropped in metal contaminated samples (Table S3) as previously Leadbetterella, Haliscomenobacter and Anaerolinea [2]. Some were previously observed [17]. Conversely, Alpha- and Betaproteobacteria classes were identified in metal contaminated sediments and soil like Flavobacterium [50], enhanced in the metal-contaminated sediments (Table S3), but only Alphapro- Methylobacter [56] and Arenimonas [57]. Some members from Sideroxydans, teobacteria were enriched at the RNA level (Table S4), implying potentially Thiobacillus or Spirochaeta have already been identified as typical slow- enhanced activity. This is in agreement with previous observations showing growing chemo-autotrophic bacteria [58–61], reinforcing the idea that the low some Alphaproteobacteria species reacting to metallic contaminants as RNA representation might be a characteristic lifestyle feature of this group. potential bioindicators in a positive (e.g. members from Sphingomonadales) [45,46], but also negative manner (e.g. members from Rhizobiales) [47]. The The groups characterized as “upcoming bacteria” were recruited in the polluted second dominant phylum in metal-contaminated sediments was Firmicutes site for their obvious tolerance to metals (Fig. 3), either under potentially active (Fig. 2), as previously described [7,12]. This group is particularly active in (FRG2) or passive (FRG3) states. Many active Firmicutes members from FRG2 metal contaminated sediments and includes mainly known metal-coping have reported metal-resistance, like Lactobacillus [49,62], Clostridium [48], bacteria from Clostridiales and Lactobacilliales [48,49]. Proteiniclasticum [63] and Turicibacter [64]. Overall, Firmicutes are prone to acquire novel beneficial genetic traits increasing their fitness [65], and some Bacteroidetes were mainly represented in the passive part of the community members are well-known for their notable natural competence capacities (e.g. (Fig. 2) and are known to be impacted by metals [17]. Although often Bacillus and Streptococcus) [66–68]. These characteristics may explain their associated with anaerobic niches, some members are affiliated to oxygenated high activity and adaptive success in MetalEurop sediments. Some active environments like representatives from Flavobacterium [50], which were Proteobacterial members were also described as harboring metal-resistant indeed detected here in the oxic upper sediment parts. In agreement with representatives, like Rhodobacter [69], Pseudomonas [13,70], Raoultella [71] previous DNA-based analysis of polluted sediments [51], Actinobacteria were and Acinetobacter [72]. The lack of phylogenetic signals inside FRG2 and more represented at MetalEurop (Fig. 2) although their RNA levels were FRG3 (Table 2) implies that their adaptive response to metal selection is either lowered, indicating some sensitivity and/or activity decrease (e.g. due to genetic parallel evolution, co-evolution/facilitation processes, or the sporulation/dormancy [52]). Most low-abundance phyla seemed significantly involvement of Horizontal Gene Transfer via broad host range mobile genetic impacted by the anthropogenic pollution at the DNA level, including elements spreading across phylogenetic barriers, as previously highlighted with Deltaproteobacteria, Gemmatimonadetes, Ignavibacteriae or Verrucomicrobia conjugative plasmids [20,73]. (Table S3). This highlights the negative effect of metals on rare members, with potential losses of important ecological functions associated with these groups “Fecal-related bacteria” (FRG4) has a pronounced RNA representation in the in aquatic ecosystems [53]. On the other hand, when comparing the average polluted site (Fig. 3), with a significant phylogenetic relatedness between the OTU RNA/DNA ratio (Fig. S3), only Actinobacteria OTUs were significantly main constituting members: Clostridium and Enterobacteriaceae (Table 2, Fig. decreased in activity in MetalEurop, confirming their sensitivity. Conversely, S2). Presence of Clostridium members in polluted and pristine sediments was Acidobacteria, Alphaproteobacteria, Betaproteobacteria, Deltaproteobacteria, previously observed [74–76], likewise for metal-resistance [48]. They also have Gemmatimonadetes, Nitrospirae, Planctomycetes and Verrucomicrobia OTUs the capacity to form endospores that can, at low temperature, produce a high unexpectedly had significantly increased activity ratios at the contaminated amount of rRNA [77] which may explain their RNA prevalence [52]. Both station. This increase could be linked to altered behavior via active resistance Clostridium and Enterobactericeae are typical bioindicators of faecal and/or stress reaction due to metal pollution, but also through changes in social contamination in water ecosystems [20,78,79]. Our results suggest that faecal- interactions between sediment microbiome members. related microorganisms could tolerate metal pollution in sediments, representing an inter-mediate environmental niche where they could potentially thrive after environmental release from wastewater treatment plants (WWTPs) 3.2. Deciphering the community tolerance and sensitivity range and/or agricultural sources. This might be related to their ability to cope with antibiotic residues often found in wastewater [20]. Indeed, bacterial resistance In order to determine microbial strategies and adaptation to anthropogenic strategies against antibiotics and metals are known to be similar and correlated metal contamination, we applied the concept of Functional Response Group, [80,81]. Resistant wastewater bacteria exposed to residual antimicrobial agents also known as groups of organisms responding similarly to environmental clues from diverse origin as WWTPs [82–84] or organic fertilizers (e.g. manure) with no phylogenetic a priori [3,20]. OTU response patterns were extracted by [85] will potentially have a selective advantage to cope with metals in means of generalized linear models under negative binomial distribution, and sediments. significance inferred by likelihood ratio tests corrected with residual deviance The group labelled as “Dominant metal sensitive bacteria” (FRG5) includes resampling with 1000 iterations (nbGLM, p < 0.05). This procedure is one of mainly prominent sediment Gammaproteobacteria members (e.g. the most reliable way to find significantly responding OTUs by minimizing the Pseudomonas sp.) with significantly reduced representation at both RNA and risk of error [54], enabling better modelling of variable inter-dependency and DNA level in the contaminated

S. Jacquiod et al. / Journal of Hazardous Materials 344 (2018) 299–307 305

site (Fig. 3). While this significant drop is likely driven by metals as previously microbes, which may explain the specific enrichment in Actinobacteria, reported [17], the significant uprising of FRG3-4 at the contaminated site could Betaproteobacteria and Firmicutes [20,98–100]. The presence of faecal-related partly explain this observation. This would indicate that dominant sediment bacteria in the metal-contaminated sediments (FRG4) likely originates from the bacteria are being challenged by both metal pollution and potential strong niche WWTP located between the two sites, and/or potentially organic fertilizers (e.g. competitors. Finally, the group labelled as “Rare metal sensitive bacteria” manure) coming from nearby agricultural fields. Furthermore, as several (FRG6) mostly contained dim archetypical oligotrophs and specialist members WWTPs are also located upstream Férin (Goeulzin and Estrée, France), it from Bacteroidetes [86], Acidobacteria [87], and Deltaproteobacteria [88], becomes more likely that the selection of these microbes could be directly which were significantly impacted by metals. Due to the rarity and significance linked to metal pollution itself, or indirectly via the local safe micro-niches of this trend, these members could be relevant bioindicator candidates of created by resistant strains. These observations relate to the concept of bacterial sediment pollution. coalescence, where microbiomes of different ecosystems meet and intermix, which may result in higher richness [101]. Furthermore, as described above, the 3.3. Significant increase in richness in the contaminated site known co-selection for antibiotic and metal resistance could also contribute in increasing the richness, as wastewater and organic fertilizers bacteria harboring Richness index and tested estimators (ACE and Chao-1 indices) highlighted the antimicrobial resistance genes may have benefited from a selective advantage higher diversity levels observed in the polluted site compared to Férin at both in metal polluted sediments. This idea is reinforced by the fact that WWTPs DNA and RNA level (Fig. 1). This was not seen previously with 454 Roche bacteria are known to potentially act as conjugative plasmid shuttles from the metagenomic sequencing due to inaccurate taxonomic affiliation of short sewage to the environment via freshwater microbial copiotrophs [20]. As a metagenomic reads [33,89,90], but also because of sequencing depth consequence, metal contaminated sediments may be seen as an intermediate limitations associated with the technology and lack of replicates [7]. This environmental hotspot contributing to the maintenance and spreading of finding does not support our initial postulate on putative diversity loss driven antibiotic/metal resistance genes located on mobile genetic elements (MGEs) by pollution, and conversely suggests that metals altered the community such as conjugative plasmids. This may be achieved either via hosting the composition in an unexpected manner. One possibility to explain this surviving and resistant microbes from effluent sources, or through freshwater observation may be linked to the toning down of some dominant sediment bacteria that acquired these genes via HGTs [20]. Gammaproteobacteria members (Fig. 2). Gammaproteobacteria are often characterized as fast growing copiotrophs [91,92], and some of them, like members from Pseudomonas, are well-characterized strong competitors (aka weed- Finally, it is important to emphasize that the higher OTU richness observed at species), prone to dominate [20,93]. Nevertheless, the effect of metals on the polluted site (Fig. 1) did not result in higher gene diversity in metagenomes Pseudomonas members (FRG5) and also competition with better-fitted groups (Fig. S2), which were mostly driven by a high functional redundancy between (FRG3 and 4) supports the idea of a microbial competition for niche sites. Nevertheless, the metaproteogenomics comparison of Gillan et al. [7] occupation. Reduction of dominant Pseudomonas members allowed novel, highlights the specific higher proportion of metal-resistance genes at the pol- metal-tolerant, slow-growing/specialist bacteria to thrive better, resulting in an luted site, as well as the presence of mobile genetic elements [7]. While no overall diversity increase. Identification of slow-growing oligotrophic direct causal evidences could be established due to differ-ences between microbes in the polluted sediments is supporting this assertion, like techniques, the enrichment of these metagenomic functions could likely be Sphingomonadales members from “upcoming bacteria” (FRG2-3), involved in associated with the specific microbial response groups selected at the biodegradation of metal associated compounds [91]. Moreover, other contaminated site (e.g. FRGs 2, 3 and 4). Nevertheless, these separate oligotrophic specialists were detected with higher RNA prevalence in the metal observations would support the presence of a metal-driven selection in favor of contaminated sediments, including Acidobacteria, Verrucomicrobia and phylogenetically close organisms sharing similar enriched genomic features Planctomycetes (Fig. S3), which is coherent with previous observations and/or mobile genetic elements. Our results tend to support this assertion, as an [94,95]. overall significant phylogenetic relatedness signal was seen at the metal contaminated sediments compared to the control (Table 2). Both observations Another phenomenon potentially participating to the increased richness at the indicate a positive selection for phylogenetically close organisms at the polluted polluted site is the extra-cellular metal precipitation ability carried by specific site, implying that the specific metagenomic enrichment in metal resistance bacteria like Pseudomonas and Bacillus members via metallophores, EPS, genes and mobile elements could be linked to this phylogenetic signal as a biogenic sulphides or calcite [96]. These bacteria may be considered as characteristic signature of selected microbes. These assertions support the idea providers of so-called “public goods”, benefiting the whole community by that resistance mechanisms could be located on narrow host-range mobile creating local safe micro-niches for sensitive bacteria to thrive [62,60]. Indeed, genetic elements, like conjugative plasmids from specific incompatibility Lactobacillus members, which are present in our sediment communities and in groups. FRG2, have this ability to bind and sequester metals, providing local safe spots where other sensitive community members could be protected [62]. Other Conflict of interests members from FRG2 and FRG4 are known to be able to ensure metal removal Authors declare no conflict of interests. like Clostridium [48], Rhodobacter [69] and Acinetobacter [97]. Our results based on FRGs strongly suggest that such local microscale facilitation Acknowledgements mechanisms could indeed enhance the survival of sensitive and/or new comers This research was funded by the ITN Marie-Curie project Train-biodiverse n◦ toward metals, contributing to an overall increase in diversity. REA 289949 (SJ), by the Fund for Scientific Research (F.R.S-FNRS) FRFC 7050357 and T.0127.14, by the Poled’ Attraction Interuniversitaire (PAI) n◦ In relation to this aspect, recruitment of new species coming from different inlet P7/25 (VC), and by the Center for Environmental and Agricultural sources along the water stream (e.g. WWTP outlets, tributary creeks, Microbiology (CREAM2) funded by The Villum Foundation (LR). All cited agricultural sources…) may also contribute to richness increase. Indeed, a funding played a role in each WWTP is actually located upstream from MetalEurop in Douai (France), releasing specific types of

306 S. Jacquiod et al. / Journal of Hazardous Materials 344 (2018) 299–307 step of the present research. Authors wish to express a deep gratitude to Luma [20] S. Jacquiod, A. Brejnrod, S.M. Morberg, W.A. Al-Soud, S.J. Sorensen, L. Riber, George Odish and all students that have attended the 2016 master course Deciphering conjugative plasmid permissiveness in wastewater microbiomes, Mol. Ecol. (2017) 1–16. “Emerging Molecular Techniques in Microbiology” provided by the Section of [21] J.T. Morton, J. Sanders, R.A. Quinn, D. Mcdonald, A. Gonzalez, Y. Vázquez-baeza, Microbiology at the University of Copenhagen for their participation to the lab et al., Balance trees reveal microbial niche differentiation, mSystems 2 (2017) work and lectures associated with this research project. Authors also wish to 00162–216. [22] N. Fortin, D. Beaumier, K. Lee, C.W. Greer, Soil washing improves the recovery of deeply thank the artist Nicolas Jacquiod who created the artwork in the total community DNA from polluted and high organic content sediments, J. graphical abstract ([email protected], website: www.nj- Microbiol. Methods 56 (2004) 181–191. gallery.com/index.php/en/). [23] A. Schöler, S. Jacquiod, G. Vestergaard, S. Schulz, M. Schloter, Analysis of soil microbial communities based on amplicon sequencing of marker genes, Biol. Fertil. Soils 53 (2017) 485–489. Supplementary data [24] R.C. Edgar, Search and clustering orders of magnitude faster than BLAST, Bioinformatics 26 (2010) 2460–2461. [25] R.C. Edgar, UPARSE: highly accurate OTU sequences from microbial Supplementary data associated with this article can be found, in the online amplicon reads, Nat. Methods 10 (2013) 996–998. version, at https://doi.org/10.1016/j.jhazmat.2017.09. 046. [26] B.J. Haas, D. Gevers, A.M. Earl, M. Feldgarden, D.V. Ward, G. Giannoukos, et al., Chimeric 16S rRNA sequence formation and detection in Sanger and 454- pyrosequenced PCR amplicons, Genome Res. 21 (2011) 494–504. [27] P.D. Schloss, S.L. Westcott, T. Ryabin, J.R. Hall, M. Hartmann, E.B. Hollister, et al., References Introducing mothur: open-source platform-independent, community-supported software for describing and comparing microbial communities, Appl. Environ. Microbiol. 75 (2009) 7537–7541. [1] M.J. Kwon, J.S. Yang, S. Lee, G. Lee, B. Ham, M.I. Boyanov, et al., Geochemical [28] Cole JR, Wang Q, Fish JA, Chai B, McGarrell DM, Sun Y, et al. Ribosomal Database Project: characteristics and microbial community composition in toxic metal-rich sediments Data and tools for high throughput rRNA analysis. Nucleic Acids Res 2014; 42: 633–642 contaminated with Au-Ag mine tailings, J. Hazard. Mater. 296 (2015) 147–157. [29] J.G. Caporaso, K. Bittinger, F.D. Bushman, T.Z. Desantis, G.L. Andersen, R. Knight, PyNAST: a flexible tool for aligning sequences to a template alignment, [2] M.P. Reis, M.F. Dias, P.S. Costa, M.P. Ávila, L.R. Leite, F.M.G. de Araújo, et al., Bioinformatics 26 (2010) 266–267. Metagenomic signatures of a tropical mining-impacted stream reveal complex [30] M.N. Price, P.S. Dehal, A.P. Arkin, Fasttree computing large minimum evolution trees microbial and metabolic networks, Chemosphere 161 (2016) 266–273. with profiles instead of a distance matrix, Mol. Biol. Evol. 26 (2009) 1641–1650.

[3] I. Nunes, S. Jacquiod, A. Brejnrod, P.E. Holm, A. Johansen, K.K. Brandt, A. [31] J.G. Caporaso, J. Kuczynski, J. Stombaugh, K. Bittinger, F.D. Bushman, E.K. Priemé, S.J. Sørensen, Coping with copper: legacy effect of copper on potential Costello, et al., QIIME allows analysis of high- throughput community sequencing activity of soil bacteria following a century of exposure, FEMS Microbiol. Ecol. data intensity normalization improves color calling in SOLiD sequencing, Nat. 92 (2016) fiw175. Methods 7 (2010) 335–336. [4] S. Nayar, B.P.L. Goh, L.M. Chou, Environmental impact of heavy metals from [32] J.G. Caporaso, C.L. Lauber, W.A. Walters, D. Berg-Lyons, C.A. Lozupone, P.J. dredged and resuspended sediments on phytoplankton and bacteria assessed in-situ Turnbaugh, et al., Global patterns of 16S rRNA diversity at a depth of millions of mesocosms, Ecotoxicol. Environ. Saf. 59 (2004) 349–369. sequences per sample, Proc. Natl. Acad. Sci. U. S. A. 108 (Suppl) (2011) 4516–4522. [5] M.Y. Sun, K.A. Dafforn, E.L. Johnston, M.V. Brown, Core sediment bacteria drive community response to anthropogenic contamination over multiple environmental [33] S. Jacquiod, J. Stenbæk, S.S. Santos, A. Winding, S.J. Sørensen, A. Priemé, gradients, Environ. Microbiol. 15 (2013) 2517–2531. Metagenomes provide valuable comparative information on soil microeukaryotes, [6] S. Roosa, R. Wattiez, E. Prygiel, L. Lesven, G. Billon, D.C. Gillan, Bacterial Res. Microbiol. 167 (2016) 436–450. metal resistance genes and metal bioavailability in contaminated sediments, [34] Ø. Hammer, P.D. Harper DaT. Ryan, Paleontological statistics software package for Environ. Pollut. 189 (2014) 143–151. education and data analysis, Palaeontol. Electron. 4 (2001) 9–18. [7] D.C. Gillan, S. Roosa, B. Kunath, G. Billon, R. Wattiez, The long-term adaptation [35] M.E. Ritchie, B. Phipson, D. Wu, Y. Hu, C.W. Law, W. Shi, G.K. Smyth, Limma of bacterial communities in metal-contaminated sediments a metaproteogenomic powers differential expression analyses for RNA-sequencing and microarray studies, study, Environ. Microbiol. 17 (2015) 1991–2005. Nucleic Acids Res. 43 (2015) e47. [8] S. Jie, M. Li, M. Gan, J. Zhu, H. Yin, X. Liu, Microbial functional genes enriched in the [36] The R Development Core Team R: A Language and Environment for Statistical Xiangjiang River sediments with heavy metal contamination, BMC Microbiol. 16 Computing, R Foundation for Statistical Computing, Vienna, Austria, 2017, ISBN 3- (2016) 179. 900051-07-0, http://www.R-project.org/ (Accessed 03 June 2013). [9] J. Mortatti, G. Meneghel De Moraes, J.-L. Probst, Heavy metal distribution in recent sediments along the Tietê River basin [São Pauro, Brazil], Geochem. J. 46 (2011) 13– [37] Y. Wang, U. Naumann, S.T. Wright, D.I. Warton, mvabund – an R package for model- 19. based analysis of multivariate abundance data, Methods Ecol. Evol. 3 (2012) 471–474. [10] S. Postel, B. Richter, Where All the Rivers Gone? In, Rivers for Life: Managing Water for People and Nature Island, Press Washington, DC, 2004, pp. 1–41. [38] P. Dixon, VEGAN, a package of R functions for community ecology, J. Veg. Sci. 14 [11] T. Ford, D. Ryan, Toxic metals in aquatic ecosystems: a microbiological (2003) 927–930. perspective, Environ. Health Perspect. 103 (1995) 25–28. [39] C.O. Webb, D.D. Ackerly, M. McPeek, M.J. Donoghue, Phylogenies and [12] I.E. Mejias Carpio, D.C. Franco, M.I. Zanoli Sato, S. Sakata, V.H. Pellizari, S. Seckler community ecology, Annu. Rev. Ecol. Syst. 33 (2002) 475–505. Ferreira Filho, D. Frigi Rodrigues, Biostimulation of metal-resistant microbial [40] E. Paradis, Shift in diversification in sister-clade comparisons: a more powerful consortium to remove zinc from contaminated environments, Sci. Total Environ. 550 test, Evolution 66 (2012) 288–295. (2016) 670–675. [41] S.W. Kembel, P.D. Cowan, M.R. Helmus, W.K. Cornwell, H. Morlon, D.D. [13] Z. Piotrowska-Seget, M. Cycon,´ J. Kozdrój, Metal-tolerant bacteria occurring in Ackerly, et al., Picante: r tools for integrating phylogenies and ecology, heavily polluted soil and mine spoil, Appl. Soil Ecol. 28 (2005) 237–246. Bioinformatics 26 (2010) 1463–1464. [14] P. Costa, M.P. Reis, M.P. Ávila, L.R. Leite, F.M.G. De Araújo, A.C.M. Salim, et al., [42] C.O. Webb, D.D. Ackerly, S.W. Kembel, Phylocom: software for the analysis of Metagenome of a microbial community inhabiting a metal-rich tropical stream phylogenetic community structure and trait evolution, Bioinformatics 18 (2008) 2098– sediment, PLoS One 10 (2015) 1–21. 2100. [15] D.C. Gillan, B. Danis, P. Pernet, G. Joly, P. Dubois, Structure of sediment- [43] X. Xu, Z. Zhang, S. Hu, Z. Ruan, J. Jiang, C. Chen, Z. Shen, Response of soil bacterial associated microbial communities along a heavy-metal contamination communities to lead and zinc pollution revealed by Illumina MiSeq sequencing gradient in the marine environment, Appl. Environ. Microbiol. 71 (2005) investigation, Environ. Sci. Pollut. Res. (2016) 1–10. 679–690. [44] L.F.V. De Oliveira, R. Margis, The source of the river as a nursery for [16] K. Laplante, N. Derome, Parallel changes in the taxonomical structure of bacterial microbial diversity, PLoS One 10 (2015) 1–11. communities exposed to a similar environmental disturbance, Ecol. Evol. 1 (2011) [45] T.A. Vishnivetskaya, J.J. Mosher, A.V. Palumbo, Z.K. Yang, M. Podar, S.D. Brown, 489–501. et al., Mercury and other heavy metals influence bacterial community structure in [17] J. Berg, K.K. Brandt, W.A. Al-Soud, P.E. Holm, L.H. Hansen, S.J. Sørensen, O. contaminated Tennessee streams, Appl. Environ. Microbiol. 77 (2011) 302–311. Nybroe, Selection for Cu-tolerant bacterial communities with altered composition but unaltered richness, via long-term cu exposure, Appl. Environ. Microbiol. 78 [46] S. Kuppusamy, P. Thavamani, M. Megharaj, K. Venkateswarlu, Y.B. Lee, R. Naidu, (2012) 7438–7446. Pyrosequencing analysis of bacterial diversity in soils contaminated long-term with [18] P.A. Sobecky, J.M. Coombs, Horizontal gene transfer in metal and radionuclide PAHs and heavy metals: implications to bioremediation, J. Hazard. Mater. 317 (2016) contaminated soils, in: M.B. Gogarten, J.P. Gogarten, L.C. Olendzenski (Eds.), (2016) 169–179. Horizontal Gene Transfer:Genomes in Flux, Humana Press, 2009, pp. 455–472. [47] J.P. Obbard, K.C. Jones, The effect of heavy metals on dinitrogen fixation by Rhizobium-white colover in a range of long-term sewage sludge amended and metal- [19] S. Roosa, C. Wauven, G. Vander Billon, S. Matthijs, R. Wattiez, D.C. Gillan, The contaminated soils, Environ. Pollut. 79 (1993) 105–112. Pseudomonas community in metal-contaminated sediments as revealed by quantitative [48] M. Alexandrino, R. Costa, A.V.M. Canário, M.C. Costa, Clostridia initiate heavy metal PCR: a link with metal bioavailability, Res. Microbiol. 165 (2014) 647–656. bioremoval in mixed sulfidogenic cultures, Environ. Sci. Technol. 48 (2014) 3378–3385.

S. Jacquiod et al. / Journal of Hazardous Materials 344 (2018) 299–307 307

[49] E. Gerbino, P. Carasi, E.E. Tymczyszyn, A. Gómez-Zavaglia, Removal unexpectedly diverse fraction of a soil bacterial community, ISME J. 9 of cadmium by Lactobacillus kefir as a protective tool against toxicity, J. (2015) 934–945. Dairy Res. 81 (2014) 280–287. [74] J.R. Matches, J. Liston, D. Curran, Clostridium [50] F. Thomas, J.H. Hehemann, E. Rebuffet, M. Czjzek, G. Michel, perfringens in the environment, Appl. Microbiol. 28 Environmental and gut Bacteroidetes the food connection, Front. (1974) 655–660. Microbiol. 2 (2011) 1–16. [75] P.S. Dobbin, J.P. Carter, C.G.S.S. Juan, M. Von Höbe, A.K. Powell, [51] A. Kaci, F. Petit, M. Fournier, S. Cécillon, D. Boust, P. Lesueur, T. D.J. Richardson, Dissimilatory Fe[III] reduction by Clostridium Berthe, Diversity of active microbial communities subjected to long-term beijerinckii isolated from freshwater sediment using Fe[III] maltol exposure to chemical contaminants along a 40-year-old sediment core, enrichment, FEMS Microbiol. Lett. 176 (1999) 131–138. [76] S. Kim, H. Jeong, J. Chun, Clostridium aestuarii sp. nov.: from Environ. Sci. Pollut. Res. 23 (2016) 4095–4110. [52] S.J. Blazewicz, R.L. Barnard, R.A. Daly, M.K. Firestone, Evaluating tidal flat sediment, Int. J. Syst. Evol. Microbiol. 57 (2007) rRNA as an indicator of microbial activity in environmental 1315–1317. [77] E. Segev, Y. Smith, S. Ben-Yehuda, RNA dynamics in aging communities: limitations and uses, ISME J. 7 (2013) 2061–2068. [53] D. Tsementzi, R. Poretsky, L.M. Rodriguez-R, C. Luo, K.T. bacterial spores, Cell 148 (2012) 139–149. Konstantinidis, Evaluation of metatranscriptomic protocols and [78] S.C. Edberg, E.W. Rice, R.J. Karlin, M.J. Allen, Escherichia coli: the application to the study of freshwater microbial communities, Environ. best biological drinking water indicator for public health protection, Microbiol. Rep. 6 (2014) 640–655. Symp. Ser. Soc. Appl. Microbiol. 88 (2000) 106S–116S. [79] S. Robles, J.M. Rodríguez, I. Granados, C.M. Guerrero, Sulfite- [54] J. Thorsen, A. Brejnrod, M. Mortensen, M.A. Rasmussen, J. Stokholm, reducing clostridia in the sediment of a high mountain lake [Laguna W.A. Al-soud, et al., Large-scale benchmarking reveals false discoveries Grande Gredos Spain] as indicator of fecal pollution, Int. Microbiol. 3 and count transformation sensitivity in 16S rRNA gene amplicon data (2000) 187–191. [80] C. Baker-Austin, M.S. Wright, R. Stepanauskas, J.V. McArthur, Co- analysis methods used in microbiome studies, Microbiome 4 (2016) 1–14. [55] F. Matyar, O. Gülnaz, G. Guzeldag, H.A. Mercimek, S. Akturk, A. Arkut, selection of antibiotic and metal resistance, Trends Microbiol. 14 (2006) M. Sumengen, Antibiotic and heavy metal resistance in Gram-negative 176–182. bacteria isolated from the Seyhan Dam Lake and Seyhan River in Turkey, [81] W. Luo, Z. Xu, L. Riber, L.H. Hansen, S.J. S??rensen, Diverse gene functions in a soil mobilome, Soil Biol. Biochem. 101 (2016) 175–183. Ann. Microbiol. 64 (2014) 1033–1040. [56] H. Zolgharnein, K. Karami, M. Mazaheri Assadi, A. Dadolahi [82] Marti, Juan Jofre, Jose Luis Balcazar, Prevalence of antibiotic resistance Sohrab, Molecular characterization and phylogenetic analyses of genes and bacterial community composition in a river influenced by a heavy metal removal bacteria from the persian gulf, wastewater treatment plant, PLoS One 8 (2013) e78906. [83] L. Rizzo, C. Manaia, C. Merlin, T. Schwartz, C. Dagot, M.C. Ploy, et Biotechnology 9 (2010) 1–8. [57] F. Chen, Z. Shi, G. Wang, Arenimonas metalli sp. nov.: isolated from an al., Urban wastewater treatment plants as hotspots for antibiotic resistant bacteria and genes spread into the environment: a review, iron mine, Int. J. Syst. Evol. Microbiol. 62 (2012) 1744–1749. [58] A.P. Harrison, B.W. Jarvis, J.L. Johnson, Heterotrophic bacteria from Sci. Total Environ. 447 (2013) 345–360. cultures of autotrophic Thiobacillus ferrooxidans: relationships as studied by means of deoxyribonucleic acid homology, J. Bacteriol. 143 (1980) [84] S.J. Kimosop, Z.M. Getenga, F. Orata, V.A. Okello, J.K. Cheruiyot, Residue levels and discharge loads of antibiotics in wastewater 448–454. [59] M. Blöthe, E.E. Roden, Composition and activity of an treatment plants [WWTPs], hospital lagoons, and rivers within autotrophic Fe[II]-oxidizing: nitrate-reducing enrichment Lake Victoria Basin, Kenya, Environ. Monit. Assess. 188 (2016) 532. culture, Appl. Environ. Microbiol. 75 (2009) 6937–6940. [60] G.M. Pumphrey, A. Ranchou-Peyruse, J.C. Spain, Cultivation- [85] H. Heuer, H. Schmitt, K. Smalla, Antibiotic resistance gene spread independent detection of autotrophic hydrogen-oxidizing bacteria by due to manure application on agricultural fields, Curr. Opin. DNA stable-isotope probing, Appl. Environ. Microbiol. 77 (2011) 4931– Microbiol. 14 (2011) 236–243. 4938. [61] M.C.Y. Lau, T.L. Kieft, O. Kuloyo, B. Linage-Alvarez, E. van Heerden, [86] S. Eichler, R. Christen, C. Höltje, J. Bötel, I. Brettar, A. Mehling, et M.R. Lindsay, et al., An oligotrophic deep-subsurface community al., Composition and dynamics of bacterial communities of a dependent on syntrophy is dominated by sulfur-driven autotrophic drinking water supply system as assessed by RNA- and composition and dynamics of bacterial communities of a drinking denitrifiers, Proc. Natl. Acad. Sci. U. S. A. 113 (2016) E7927–E7936. [62] M. Monachese, J.P. Burton, G. Reid, Bioremediation and human tolerance water supply system as assessed by RNA- and DNA-Based 16S to heavy metals through microbial processes: a potential for probiotics? rRNA gene fingerprinting, Appl. Environ. Microbiol. 72 (2006) 1858–1872. Appl. Environ. Microbiol. 78 (2012) 6466–6474. [63] Y. Ren, J. Niu, W. Huang, D. Peng, Y. Xiao, X. Zhang, et al., [87] N. Fierer, M.A. Bradford, R.B. Jackson, Toward an ecological Comparison of microbial taxonomic and functional shift pattern classification of soil bacteria, Ecology 88 (2007) 1354–1364. [88] W. Ye, X. Liu, S. Lin, J. Tan, J. Pan, D. Li, H. Yang, The vertical along contamination gradient, BMC Microbiol. 16 (2016) 1–9. [64] J. Breton, C. Daniel, C. Vignal, M. Body-Malapel, A. Garat, C. Ple, B. distribution of bacterial and archaeal communities in the water and Foligne, Does oral exposure to cadmium and lead mediate susceptibility sediment of Lake Taihu, FEMS Microbiol. Ecol. 70 (2009) 263–276. to colitis? the dark-and-bright sides of heavy metals in gut ecology, Sci. [89] R. Ranjan, A. Rani, A. Metwally, H.S. McGee, D.L. Perkins, Rep. 6 (2016) 19200. Analysis of the microbiome: advantages of whole genome shotgun versus 16S amplicon sequencing, Biochem. Biophys. Res. [65] V.F. Lanza, A.P. Tedim, J.L. Martínez, F. Baquero, T.M. Coque, V.F. Commun. 469 (2016) 967–977. Lanza, et al., The plasmidome of firmicutes: impact on the emergence and [90] N. Shah, H. Tang, T.G. Doak, Y. Ye, Comparing bacterial communities the spread of resistance to antimicrobials, Microbiol. Spectr. 3 (2015) 1– inferred from 16S rRNA gene sequencing and shotgun metagenomics, Pac. Symp. Biocomput. (2011) 165–176. 37. [66] P.J. Johnsen, D. Dubnau, B.R. Levin, Episodic selection and the [91] F.M. Lauro, D. McDougald, T. Thomas, T.J. Williams, S. Egan, S. Rice, maintenance of competence and natural transformation in Bacillus et al., The genomic basis of trophic strategy in marine bacteria, Proc. Natl. Acad. Sci. U. subtilis, Genetics 181 (2009) 1521–1533. [67] B.A. Evans, D.E. Rozen, Significant variation in transformation S. A. 106 (2009) 15527–15533. [92] S. Jacquiod, L. Franqueville, S. Cécillon, T.M. Vogel, P. Simonet, Soil frequency in Streptococcus pneumoniae, ISME J. 7 (2013) 791–799. [68] J.C. Mell, R.J. Redfield, Natural competence and the evolution of DNA bacterial community shifts after Chitin enrichment: an integrative metagenomic approach, PLoS One 8 (2013) 1–13. uptake specificity, J. Bacteriol. 196 (2014) 1471–1483. [69] L. Giotta, A. Agostiano, F. Italiano, F. Milano, M. Trotta, Heavy [93] J.A. Cray, A.N.W. Bell, P. Bhaganna, A.Y. Mswaka, D.J. Timson, J.E. metal ion influence on the photosynthetic growth of Rhodobacter Hallsworth, The biology of habitat dominance; can microbes behave as weeds? Microbiol. Biotechnol. 6 (2013) 453–492. sphaeroides, Chemosphere 62 (2006) 1490–1499. [70] S. O’Brien, D.J. Hodgson, A. Buckling, Social evolution of toxic [94] G.T. Bergmann, S.T. Bates, K.G. Eilers, C.L. Lauber, G.J. Caporaso, metal bioremediation in Pseudomonas aeruginosa, Proc. Biol. Sci. W.A. Walters, et al., The under-recognized dominance of Verrucomicrobia in soil bacterial communities, Soil Biol. Biochem. 43 281 (2014) 20140858. [71] S. Koc, B. Kabatas, B. Icgen, Multidrug and heavy metal-resistant (2011) 1450–1455. Raoultella planticola isolated from surface water, Bull. Environ. [95] C. Jenkins, V. Kedar, J.A. Fuerst, Gene discovery within the planctomycete division of the domain Bacteria using sequence tags Contam. Toxicol. 91 (2013) 177–183. [72] M.H. El-Sayed, Multiple heavy metal and antibiotic resistance of from genomic DNA libraries, Genome Biol. 3 (2002) acinetobacter baumannii strain HAF –13 isolated from industrial (RESEARCH0031). [96] D.C. Gillan, Metal resistance systems in cultivated bacteria: are they effluents, Am. J. Microbiol. Res. 4 (2016) 26–36. [73] U. Klümper, L. Riber, A. Dechesne, A. Sannazzarro, L.H. Hansen, S.J. found in complex communities? Curr. Opin. Biotechnol. 38 (2016) Sørensen, B.F. Smets, Broad host range plasmids can invade an 123–130.

[97] D. Abdel-el-haleem, Minireview Acinetobacter: [99] S.J. McIlroy, A.M. Saunders, M. Albertsen, M. Nierychlo, B. environmental and biotechnological applications, Afr. J. McIlroy, A.A. Hansen, et al., MiDAS: the field guide to the Biotechnol. 2 (2003) 71–74. microbes of activated sludge, Database 2015 (2015) bav062. [98] S. Atashgahi, R. Aydin, M.R. Dimitrov, D. Sipkema, K. Hamonts, L. [100] A. Cydzik-Kwiatkowska, M. Zielinska, Bacterial communities in Lahti, et al., Impact of a wastewater treatment plant on microbial full-scale wastewater treatment systems, World J. Microbiol. community composition and function in a hyporheic zone of a eutrophic Biotechnol. 32 (2016) 1–8. river, Sci. Rep. 5 (2015) 17284. [101] M.C. Rillig, J. Antonovics, T. Caruso, A. Lehmann, J.R. Powell, S.D. Veresoglou,, E.Verbruggen, Interchange of entire communities: microbial community coalescence, Trends Ecol. Evol. 30 (2015) 470– 476.

Supporting Figure File

Samuel Jacquiod, Valentine Cyriaque, Leise Riber, Waleed Abu Al-soud, David C. Gillan, Ruddy Wattiez, Søren J. Sørensen. “Long-term industrial metal contamination unexpectedly shaped diversity and activity response of sediment microbiome”.

Journal of Hazardous Materials (2017)

Figure S1: Rarefaction curves of individual 16S rRNA gene amplicon samples for the Férin DNA (dF) and RNA (cF) levels and for the MetalEurop sites DNA (dM) and RNA (cM) levels. Despite count fluctuation, most microbiome profiles displayed the typical asymptotic plateau trend, indicating a sufficient sequencing coverage for appropriate characterization of the core diversity (Figure S1). This was particularly true for DNA-based profiles compared to their RNA counterparts, which were less saturated

Figure S2: Venn diagram displaying the OTU richness distribution and the metagenomic functions diversity in both sites. Panel A) shows the 16S rRNA gene OTU distribution from DNA and RNA datasets in each sites (Férin and MetalEurop), while panel B) displays the repartition of shared and unique genetic functional features in both sites from metagenomic data. The sums of all 16S rRNA gene replicates were resampled to the same level to account for unevenness (66,000 counts per dataset). Metagenome profiles were also resampled to 160,000 counts for similar reasons.

Figure S3: Average RNA/DNA ratio from the 16S rRNA amplicon sequencing per OTU in phylogenetic groups in each site (±SEM). Values below or above 1 correspond to an under- and over-representation of RNA compared to DNA. Significant differences between Férin and MetalEurop were identified with one-sided Student test (*** if p < 0.001; ** if p < 0.01; * if p < 0.05).

Figure S4: Generalized heatmap of significantly responding OTUs across experimental design. Significance of OTU response patterns were extracted by means of generalized linear models under negative binomial distribution and likelihood ratio tests corrected with residual deviance resampling (nbGLM, p < 0.05, 1,000 iterations). Functional response grouping was done with hierarchical clustering based on center-scaling of abundance. The six groups are separated by green, pink, orange, blue, yellow and purple colors from FRG1 to FRG6.

Figure S5: Statistical verification and validation of the abundance driven pattern identified with the three FRGs by means of constrained PCA followed by a random null model Monte-Carlo simulation implementing 100,000 group permutations (p < 1.0E-07).

Supporting Table File

Table S1: Total metal concentrations (mean±SD; n=4) in Férin and MetalEurop sediments. Adapted from Gillan et al, 2015.

Férin MetalEurop Al (g/kg) 17.6±3.0 26.2±0.7 As (mg/kg) 2.8±0.3 21.0±0.9 Cd (mg/kg) 1.3±0.03 38.1±0.5 Co (mg/kg) 5.8±0.01 8.8±0.3 Cr (mg/kg) 56.2±1.7 107.4±1.9 Cu (mg/kg) 13.7±0.4 100.0±0.8 Fe (g/kg) 12.0±0.4 20.6±0.3 Mn (mg/kg) 293.5±7.6 547.9±1.4 Ni (mg/kg) 15.2±0.7 25.5±0.4 Pb (mg/kg) 111.6±0.8 913.8±11 V (mg/kg) 35.9±0.4 62.1±0.6 Zn (mg/kg) 348.5±6.7 3218.5±69

Table S2: Summary table of 16S rRNA gene amplicon profiles obtained. Samples below 2,000 counts were not included into the analysis (dF6 and dM4).

Sample Sort Station Core Counts Suitability dF1 DNA Férin A 37047 YES dF2 DNA Férin A 18269 YES dF3 DNA Férin B 16615 YES dF4 DNA Férin B 14643 YES dF5 DNA Férin C 8966 YES dF6 DNA Férin C 1549 NO dM1 DNA Metal D 20705 YES dM2 DNA Metal D 14889 YES dM3 DNA Metal E 16102 YES dM4 DNA Metal E 282 NO dM5 DNA Metal F 5016 YES dM6 DNA Metal F 10216 YES rF1 RNA Férin A 6618 YES rF2 RNA Férin A 5546 YES rF3 RNA Férin B 16897 YES rF4 RNA Férin B 21369 YES rF5 RNA Férin C 20073 YES rF6 RNA Férin C 16895 YES rM1 RNA Metal D 20373 YES rM2 RNA Metal D 17970 YES rM3 RNA Metal E 15475 YES rM4 RNA Metal E 14027 YES rM5 RNA Metal F 35665 YES rM6 RNA Metal F 20091 YES

Table S3: Phylogenetic composition of the prokaryotic community of sediments between Férin and MetalEurop. The table shows average relative abundance (±SE) distribution of phylogenetic groups for DNA (n = 5) and RNA (n = 6) profiles. Statistical significance was inferred by ANOVA with false discovery rate post-hoc multiple correction test (FDR, p < 0.05). Letter “a” and “b”, respectively, indicates the lowest and highest values in case of significance.

DNA level RNA level

Phylogenetic group Férin (%) Metal (%) FDR p Férin (%) Metal (%) FDR p Acidobacteria 0.08 ±0.02b 0.002 ±0.002a 8.08E-03 0.01 ±0.004 0.01 ±0.01 0.500 Actinobacteria 0.98 ±0.05a 3.93 ±0.41b 3.16E-03 0.59 ±0.12 0.81 ±0.12 0.469 Alphaproteobacteria 0.91 ±0.02a 1.41 ±0.06b 2.85E-03 0.5 ±0.08a 1.03 ±0.12b 0.031 Bacteroidetes 2.7 ±0.06 2.79 ±0.14 0.677 0.3 ±0.07 0.26 ±0.05 0.783 Betaproteobacteria 3.23 ±0.2a 5.04 ±0.41b 0.018 1.28 ±0.27 0.84 ±0.13 0.381 Caldiserica 0.002 ±0.001 0.01 ±0.002 0.380 0.001 ±0.001 0.004 ±0.002 0.381 Chloroflexi 4.6 ±1.02 4.38 ±0.09 0.856 2.98 ±0.63 2.27 ±0.55 0.608 Crenarchaeota 0.03 ±0 0.02 ±0.01 0.475 0.01 ±0 0.004 ±0.002 0.711 Deferribacteres 0.02 ±0.005 0.01 ±0.01 0.394 0 0 1.043 Deinococcus-Thermus 0.002 ±0.002 0 0.416 0 0 1.000 Deltaproteobacteria 1.02 ±0.11b 0.5 ±0.05a 0.013 0.26 ±0.05 0.33 ±0.08 0.678 Epsilonproteobacteria 0.01 ±0.004a 0.09 ±0.02b 9.54E-03 0.001 ±0.001a 0.04 ±0.01b 2.10E-03 Euryarchaeota 1.14 ±0.09 1.3 ±0.24 0.679 1.1 ±0.37 4 ±2.14 0.464 Firmicutes 3.88 ±0.65a 9.24 ±0.56b 4.09E-03 9.45 ±2.64a 36.5 ±2.03b 3.36E-04 Fusobacteria 0.01 ±0.003 0.001 ±0.001 0.172 0 0.001 ±0.001 0.545 Gammaproteobacteria 78.81 ±1.26b 69.57 ±1.06a 5.30E-03 81.95 ±3.07b 52.69 ±3.1a 9.50E-04 Gemmatimonadetes 0.2 ±0.02b 0.08 ±0.03a 0.025 0.13 ±0.04 0.12 ±0.03 0.961 Ignavibacteriae 0.52 ±0.04b 0.35 ±0.03a 0.018 0.22 ±0.05 0.12 ±0.03 0.350 Nitrospirae 0.25 ±0.05b 0.03 ±0.01a 8.33E-03 0.16 ±0.04 0.03 ±0.01 0.030 Other Proteobacteria 0.04 ±0.003b 0.01 ±0.004a 4.66E-03 0.01 ±0.01 0.002 ±0.001 0.341 Planctomycetes 1.12 ±0.11b 0.56 ±0.06a 8.82E-03 0.67 ±0.14 0.64 ±0.15 0.952 Spirochaetes 0.13 ±0.01a 0.6 ±0.05b 2.51E-03 0.03 ±0.01b 0.16 ±0.02a 2.70E-03 Unclassified Bacteria 0.18 ±0.02b 0.07 ±0.01a 0.013 0.29 ±0.08 0.09 ±0.02 0.145 Verrucomicrobia 0.16 ±0.03b 0.01 ±0.005a 7.72E-03 0.07 ±0.01 0.05 ±0.01 0.570

Table S4: Phylogenetic composition of the DNA and RNA prokaryotic community profiles in each site. The table shows average relative abundance (±SE) distribution of phylogenetic groups for DNA (n = 5) and RNA (n = 6) profiles. Statistical significance was inferred by ANOVA with false discovery rate post-hoc multiple correction test (FDR, p < 0.05). Letter “a” and “b”, respectively, indicates the lowest and highest values in case of significance.

Férin station MetalEurop station

Phylogenetic group DNA (%) RNA (%) FDR p DNA (%) RNA (%) FDR p Acidobacteria 0.08 ±0.02b 0.01 ±0.004a 6.10E-03 0.002 ±0.002 0.01 ±0.01 0.219 Actinobacteria 0.98 ±0.05b 0.59 ±0.12a 0.043 3.93 ±0.41b 0.81 ±0.12a 6.33E-04 Alphaproteobacteria 0.91 ±0.02b 0.5 ±0.08a 6.89E-03 1.41 ±0.06 1.03 ±0.12 0.107 Bacteroidetes 2.7 ±0.06a 0.3 ±0.07b 4.07E-08 2.79 ±0.14b 0.26 ±0.05a 2.00E-06 Betaproteobacteria 3.23 ±0.2b 1.28 ±0.27a 2.77E-03 5.04 ±0.41b 0.84 ±0.13a 8.83E-05 Caldiserica 0.002 ±0.001 0.001 ±0.001 0.433 0.01 ±0.002 0.004 ±0.002 0.776 Chloroflexi 4.6 ±1.02 2.98 ±0.63 0.286 4.38 ±0.09b 2.27 ±0.55a 0.035 Crenarchaeota 0.03 ±0b 0.01 ±0a 0.012 0.02 ±0.01 0.004 ±0.002 0.116 Deferribacteres 0.02 ±0.005b 0a 6.57E-03 0.01 ±0.01 0 0.377 Deinococcus-Thermus 0.002 ±0.002 0 0.339 0 0 1.000 Deltaproteobacteria 1.02 ±0.11b 0.26 ±0.05a 2.07E-03 0.5 ±0.05 0.33 ±0.08 0.293 Epsilonproteobacteria 0.01 ±0.004 0.001 ±0.001 0.246 0.09 ±0.02 0.04 ±0.01 0.111 Euryarchaeota 1.14 ±0.09 1.1 ±0.37 0.934 1.3 ±0.24 4 ±2.14 0.492 Firmicutes 3.88 ±0.65 9.45 ±2.64 0.154 9.24 ±0.56a 36.5 ±2.03b 2.30E-05 Fusobacteria 0.01 ±0.003 0 0.070 0.001 ±0.001 0.001 ±0.001 1.020 Gammaproteobacteria 78.81 ±1.26 81.95 ±3.07 0.427 69.57 ±1.06b 52.69 ±3.1a 6.88E-03 Gemmatimonadetes 0.2 ±0.02 0.13 ±0.04 0.226 0.08 ±0.03 0.12 ±0.03 0.647 Ignavibacteriae 0.52 ±0.04b 0.22 ±0.05a 5.12E-03 0.35 ±0.03b 0.12 ±0.03a 4.11E-03 Nitrospirae 0.25 ±0.05 0.16 ±0.04 0.249 0.03 ±0.01 0.03 ±0.01 0.948 Other Proteobacteria 0.04 ±0.003 0.01 ±0.01 0.072 0.01 ±0.004 0.002 ±0.001 0.809 Planctomycetes 1.12 ±0.11b 0.67 ±0.14a 1.52E-05 0.56 ±0.06a 0.64 ±0.15b 5.10E-04 Spirochaetes 0.13 ±0.01 0.03 ±0.01 0.332 0.6 ±0.05 0.16 ±0.02 0.648 Unclassified Bacteria 0.18 ±0.02a 0.29 ±0.08b 9.12E-03 0.07 ±0.01 0.09 ±0.02 0.652 Verrucomicrobia 0.16 ±0.03b 0.07 ±0.01a 0.037 0.01 ±0.005 0.05 ±0.01 0.105

Chapter 2: Metal-induced bacterial interactions promote diversity in river-sediment microbiomes

Metal-induced bacterial interactions promote diversity in river-sediment microbiomes

1 # 1,2 3 4 1 Valentine Cyriaque * , Augustin Géron *, Gabriel Billon , Joseph Nesme , David C. Gillan , Søren J. Sørensen4 and Ruddy Wattiez1 1. Proteomics and Microbiology Lab, Research Institute for Biosciences, UMONS, Mons, Belgium 2. Division of Biological and Environmental Sciences, Faculty of Natural Sciences, University of Stirling, Stirling, UK. 3. Univ. Lille, CNRS, UMR 8516 – LASIR – Laboratoire de Spectrochimie Infra-rouge et Raman, F- 59000 Lille, France 4. Section of Microbiology, Department of Biology, University of Copenhagen, Denmark.

*Both authors have contributed equally to this work and share first authorship.

#Corresponding author: [email protected]; Tel.: +32 (0)65 37 33 19; Proteomics and microbiology laboratory, Research Institute for Biosciences, UMONS, 20 place du parc, Mons, Belgium Abstract Background. Anthropogenic metal contamination results in long-term environmental selective pressure with unclear impacts on bacterial communities, key players in ecosystem functioning. Since metal contamination poses serious toxicity and bioaccumulation issues, assessing their impact on environmental microbiomes is important to respond to current environmental and health issues. Despite elevated metal concentrations, river sedimentary microbiome near the MetalEurop foundry (France) show unexpected higher diversity compared to upstream control site. In this work, a follow-up of the microbial community assembly during a metal contamination event was performed in microcosms with periodic renewal of the supernatant river water. Results. Sediments of the control site were gradually exposed to a mixture of metals (Cd, Cu, Pb and Zn), in order to reach similar concentrations to MetalEurop sediments. Illumina sequencing analysis of 16S rRNA gene amplicons was performed. Metal resistance genes czcA and pbrA, as well as IncP plasmid content were assessed by quantitative PCR. The outcomes of this study support previous in-situ observations showing that metals act as community assembly managers, increasing diversity. Conclusion. This work revealed progressive adaptation of the sediment microbiome through the selection of different metal-resistance mechanisms and cross-species interactions involving public good providing bacteria co-occurring with the rest of the community.

Keywords: river sediment microbiome, community assembly, metal, diversity, facilitator bacteria

1.Background Bacteria are key players in ecosystems as they are involved in all biogeochemical cycles

[1]. In river ecosystems, bacteria are generally very active and intensively colonize muddy sediments [2, 3]. The river microbial network is established as a directional linear branched structure shaped by the riverine flow that accelerates the dispersal of microorganisms from upstream [4], enriching the downstream sediments [4–8]. The branched structure of the river ecosystem brings bacteria from surrounding lands, including urban and industrial areas, wastewater treatment plants (WWTPs) and agricultural sources that also supply soluble components [7]. As metals react with particles (oxides, clays, organic matter and sulfides) through adsorption and (co)precipitation, sediments are an efficient sink for inorganic pollutants such as metals [7, 9, 10].

The use of metals in anthropogenic activities increased their dissemination in the environment [11, 12], including in river ecosystems [13]. While some metals are essential nutrients, all have harmful toxic effects at high concentrations. Among prokaryotes, the most widespread metal resistance systems found in in-situ bacterial communities are in the periplasm or anchored in the cytoplasmic membrane [14] and are largely represented by efflux pumps. In addition to the direct effects of metal contamination on biological systems, metal-induced stresses co-select for antibiotic resistance mechanisms, for example by co-selection of mechanisms involved in efflux systems [15–18]. Furthermore, public good providing bacteria were shown to provide interspecific benefits in a complex community [19]. These bacteria, also called facilitators, detoxify sediments (e.g. via metal precipitation and sequestration) and provide safe micro-niches for metal-sensitive bacteria (e.g. biofilm formation) [20]. Metal contamination was found to maintain [1, 21], decrease [22] or increase microbiome richness and evenness [7]. These contrasting observations might be linked to both the site properties

(e.g., initial composition of the microbiome, the nature of contaminants, exposure times or local geo-physicochemical parameters).

The Deûle River (Northern France) flows next to the former foundry MetalEurop where sediments are up to 30-fold more contaminated with metals than an upstream control site (the

Sensée canal, Férin) [1]. Interestingly, MetalEurop sediments exhibited a highly diverse microbiome compared to the control site [1, 7]. Jacquiod and colleagues suggested that metals may play a key role in the assembly of in-situ microbial communities through the coalescence

(i.e. the gathering of microbiomes from different ecosystems [23] ) of indigenous bacteria and bacteria from different upstream environmental sources (e.g. WWTPs or agricultural sources).

Authors suggested that metals increased diversity by (i) preventing the development of metal- sensitive opportunists and (ii) selecting metal-resistant bacteria providing public-goods for the rest of the community. In addition, Horizontal Gene Transfer (HGT) has been suggested to be involved in metal-resistance gene dispersal in the community [7]. If plasmids present in the environment, especially IncP plasmids, have been shown to carry metal resistance genes [24,

25], the impact of metals on their persistence and dispersion in the environment is not clear.

In the present study, we incubated control sediments in microcosm for more than 6 months, some of the microcosms being exposed to metals that gradually reached those found in nearby contaminated site MetalEurop: Cd (~ 40 mg/kg), Cu (~100 mg/kg), Pb (~900 mg/kg) and Zn (~3000mg/kg)[1, 7]. The overlaying water was renewed periodically with fresh river water to induce community coalescence phenomenon such as observed at natural conditions [6,

7]. We performed a follow-up of the taxonomic and functional evolution of river sedimentary microbiomes from the Sensée canal (Férin, France) in the presence of metal concentrations.

The community assemblage in sediment was monitored over time using 16S rRNA gene amplicon sequencing and quantitative PCR to follow the metal resistance genes czcA and pbrA, encoding metal efflux pumps, but also, the evolution of the IncP plasmid family.

2. Results

2.1. The α-diversity of river-like metal-impacted microcosms remains high over time. Species richness and Shannon’s diversity index decreased slowly during the 6.5 months of incubation (Figure 1). Both indices were significantly higher in the control microcosms than in the metal-treated microcosms after 15 days of incubation. Richness fluctuated at the same level in all microcosms, while after 3.5 months, Shannon index was significantly higher in the metal-treated microcosms than in the controls (Figure 1).

Figure 1: α-diversity (OTU richness and Shannon indexes ± SEM) obtained from 16S rRNA gene amplicon sequencing carried out from Férin sediment on the day of sampling (in- situ) and from control and metal-treated microcosm sediments (n=4). *: p-value<0.05; **: p- value<0.01; ***:p-value<0.001.

2.2. Community succession in control and metal-impacted microcosms PERMANOVA revealed significant changes in ß-diversity over time (Table S3). Time, metals and the interaction of both significantly affected the sediment microbial structure.

Phylogenetic indices changed over time in both control and metal-treated microcosms: (i) At the start of the experiment, NRI index of the metal-treated microcosms increased and then fluctuated around 0. In control microcosms NRI index immediately decreased and reached a

value of -2.20 after 4.5 months but came back to initial values after 6.5 months (Figure 2A);

(ii) β-NRI fluctuated around 0 in the metal-treated microcosms and got close to 2 in control microcosms before coming back to initial values (Figure 2B); (iii) NTI was high (>2) and relatively stable over time in both conditions (Figure 2C) and; (iv) β-NTI index started at –1 at the beginning of the experiment and progressively increased to over 0 after 3.5 months (Figure

2D).

Figure 2: Changes in the NRI (net relatedness index, A), β-NRI (B), NTI (nearest taxon index,

C) and β-NTI (D), over time determined from 16S rRNA gene amplicon sequencing carried out on DNA extracted from Férin in-situ sediment on the day of sampling (in-situ) and from control and metal-treated microcosm sediments using weighted OTU representation and 1 000 randomizations (n=4).

Before the first 3.5 months, there were no significant differences in phyla abundance between control and metal-treated microcosms (Table S4). From 3.5 months, significant

differences were observed for several phyla, including α-Proteobacteria, β-Proteobacteria,

Cyanobacteria and δ-Proteobacteria, Firmicutes, Euryarchaeota, Chloroflexi or

Saccharibacteria. In the control microcosms, Cyanobacteria, Euryarchaeota and

Saccharibacteria were found to be enriched.

A total of 600 OTUs responded significantly to metal stress (p < 0.05) and were classified into 5 TRGs (Figure S1) depending on their abundance in control and metal-treated microcosms over time. TRGs were validated with Monte Carlo simulation (Figure S2). The TRG “A” contained OTUs that were abundant in the control and rare in the metal-treated microcosms

(Figures S3 & 3). The four TRGs “B” (Groups B1–B4) contained OTUs that were more abundant in the metal-treated microcosms and rare in the controls. Group B1 contained OTUs showing an increased proportion just after the metal addition (i.e., at 0.5 and 1.5 months;

Figures S3 & 3). The OTUs in Groups B2 and B3 showed high proportions successively in the middle of the incubation period, and the OTUs in Group B4 showed high proportions at the end of the experiment (Figures S3 and 3).

At the order level, a peak of methanogenic archaea was observed after 2.5 months in the control microcosms (Methanosarcinales) and in smaller proportion, after 4.5 months

(Methanomicrobia and Methanomassillicoccus), in the metal-treated microcosms. Bacteria related to Candidatus Saccharibacteria clearly proliferated in the control microcosms but were almost absent in the metal-treated microcosms (Figure 3). Firmicutes were found in all B groups and were mainly represented by Clostridiales (B1 (Clostridium and Desulfosporinus), B2, B4

(Ruminococcaceae)), Negativicutes (B2 and B4; Veillonellaceae) and Selemonadales (B3, B4).

Within α-Proteobacteria, Rhodospirillales (Roseomonas) was detected in group B1,

Sphingomonadales, Rhizobiales and Rhodobacterales appeared in the group B2 and

Caulobacterales (Brevundimonas) arrived lately in group B4. Among β-proteobacteria,

Burkohlderiales (Aquabacterium, Simplicispira), Rhodocyclales (Dechloromonas, Zoogloea)

and Methylophilales were observed in group B1 and Rhodocyclales in the group B3.

Legionellales (Legionella) from γ-proteobacteria was found in group B1. Sulphate Reducing

Bacteria (SRB) from δ-Proteobacteria (Desulfobulbus, Desulfovibrio, Geobacter) were detected in group B3 and Anaerolineales (Chloroflexi, Leptolinea, Longilinea, Anaerolinea) were found in groups B3 and B4 (Figure 3). Finally, Terrimonas (Sphingobacteriales,

Bacteroidetes), Prosthecobacter (Verrucomicrobiales, Verrucomicrobia) and Geothrix

(Holophagales, Acidobacteria) were found in groups B1, B2 and B3 respectively.

Figure 3: Relative Abundances (percentage of read counts) and their corresponding taxonomical families composing TRGs in the controls (left) and in metal impacted microcosms

(right). OTUs significantly responding to metal contamination over time were highlighted using

a nbGLM (p<0.05) and TRGs were identified by displaying those OTUs in a heatmap

(Supplemental figure S1).

2.3. Dynamic occurrence network reveals links between hub species and facilitator bacteria. The eLSA approach was used to identify simultaneous and delayed OTU occurrences within the microbiome. The clustering coefficients (i.e. degree to which nodes grouped together) were

0.23 and 0.28 and the average number of neighbors was 2.9 and 8.3 for the control and the metal-treated microcosms respectively. In control microcosms, 3 hub OTUs (i.e. OTU 356

(Hyphomicrobiaceae), OTU 308 (Chromatiales) and OTU 358 (Verrucomicrobia from

Subdivision 3)) were found to be highly connected with the rest of the community (Figure S4A).

In metal-treated microcosms, OTU 339 (Xanthobacter), OTU 46 (unclassified -

Proteobacterium), OTU 207 (Verrucomicrobia from Subdivision 3) and OTUs 90 and 131

(unclassified bacteria, TRG B4) (Figure S4B), were identified as hub OTUs (Figure 4). Hub

OTUs displayed negative edges with Nitrospira in all microcosms and with Ignavibacteriae in metal-impacted microcosms. In addition, Xanthobacter and OTU 131 displayed negative edges with OTU 276 (Rhodocyclaceae), OTU 288 (Caldilinea), OUT 3738 (Bacteroidales) and OTU

93 (Ignavibacterium) (TRG A) (Figure S4B). Positive edges were observed between hub OTUs and facilitators bacteria. Links were found between (i) hub OTU 131 and OTUs 4886, 4189

(Zoogloea) and 111 (Dechloromonas, TRG B1), (ii) hub OTU 46 and OTU 9 (Zoogloea), (iii) hub OTU 207 and OTU 107 (Dechloromonas) and (iv) hub OTUs 339 & 131 and OTU 64

(Desulfosporinus). Anaerolineaceae (TRG B3) was also found to positively link with the rest of the community (Figure 4A and S4B), including OTU 266 (Methanobacterium), OTU 533

(Methanolinea) and members of -, -Proteobacteria and Bacteroidetes (Figure 4). In addition, other SRBs such as OTU 27 (Desulfobulbus), OTU 122 (Desulfovibrio) and OTU 55

(Geobacter, TRG B1) co-occurred or were preceded by most hub OTUs and Anaerolineaceae

(Figure 4A). Finally, OTU169 (Veillonellaceae, TRG B4) was linked with OTU 90

(Anaerolineaceae, TRG B4), OTU 207 (Verrucomicrobia) and Xanthobacter as well as other bacteria from B4 TRG (i.e. OTU 130, Rhizobacter), to B1 TRG (i.e. OTU 278, Terrimonas) and TRG unclassified bacteria (Figure 4B).

Figure 4: Network clusters gathering OTUs of interest extracted from global network

(Supplemental Figure 4) built using most significant extended local similarity scores (LS>0.8, p-value < 0.05, n=4). Green lines represent positive association of OTUs, and red lines represent negative associations. Directed edge (arrows) represents a local similarity with a time delay, representing succession of OTUs. Cluster A represents privileged edges between (i)

Anaerolineaceae and (ii) hub OTUs 90, 131,46,339 (Xanthobacter) and 207 (Verrucomicrobia, subdivision3); (iii) SRBs (Desulfobulbus (OTU 27), Desulfovibrio (OTU 122) and Geobacter

(OUT 55)), (iv) Zoogloea (OTUs 4189 and 4886) and Dechloromonas (OTUs 107 and 111), and (v) methanogen Euryarchaeota (Methanobacterium (OTU 266) and Methanolinea (OTU

533)). Cluster B represents privileged edges between the VeillonellaceaeOTU 169 and other members of the community.

2.4. IncP plasmids and metal-resistant genes content assessed by Q-PCR Metals induced IncP plasmid copy depletion over the 2.5 first months, while enrichment was observed in control microcosms. Thereafter, relative abundance stabilized until the end of the experiment in both conditions (Figure 5). No significant differences were observed in czcA relative abundance in both conditions, excepted after 4.5 months, where it significantly increased in the metal-treated microcosms (Figure 5). The pbrA level was always significantly higher in the metal-treated microcosms than in the control microcosms (Figure 5).

Figure 5: Changes in the relative amounts of czcA (A), pbrA (B) and IncP oriT (C) gene

ΔCt target (T0 – Tx) ΔCt 16S (T0 – Tx) sequences ((EincP) / (E16S rRNA) rRNA ) in each condition (±SEM)

(Metal vs Control). The significance was calculated with a paired t-test. Different letters indicate significant differences (p-value <0.01) between the two groups.

3. Discussion

3.1. The community diversity in metal-impacted river-like microcosms Metals affected the community structure after 15 days of exposure by causing a significant decrease of alpha-diversity indices. However, after 3.5 months, this trend was reversed, confirming previous in-situ observations [7]. Interestingly, in a preliminary similar experiment, where sterilized water renewal was used, Shannon index in metal-contaminated microcosms remained lower than in control condition (Figure S5). In the present research, the evolution of phylogenetic indices over time was assessed assuming that: (i) overdispersion

(decreased NRI or NTI [26]) would result from diversification of the community in line with the community coalescence process (i.e., the constant arrival of new bacteria by water renewal) and (ii) increased relatedness (increased NRI or NTI [26]) would result from environmental filtering (i.e., killing or inhibition of some bacteria by metals). In control microcosms, the NRI index decreased under the threshold value (-2), highlighting a phylogenetic overdispersion of the community. This may be explained by a community coalescence process with novel bacteria introduced during water renewal. In metal-treated microcosms and at the very start of the experiment, NRI was found to increase, showing initial environmental filtering most likely for metal-tolerant bacteria, but stabilized around 0 in the following months. Despite metals inhibited overdispersion, the coalescence process may indeed avoid strong environmental filtering. In both control and metal-treated microcosms, β-NTI was first negative but increased over 0 after 3.5 months showing no metal impact in the selection of the nearest phylogenetic neighbors. These results highlighted that the periodic arrival of new bacteria in a metal- contaminated environment helped the microbiome to persist, leading to a fast resilience and an unexpectedly high diversity.

100

3.2. Mechanisms of resilience in metal-impacted microbiome Previous studies suggested (i) broad-host-range plasmids [7, 27][1, 7], and (ii) bacterial cooperation via facilitator bacteria [7], to play an important role in the resilience observed in long-term stressed communities, such as those in metal-impacted river sediments. Broad host range plasmids of the IncP family have been found in areas contaminated with manure, wastewater of industrial origin or river sediments and often carry catabolic, antibiotic- and metal-resistance genes [28, 29]. In addition, IncP plasmids were shown to be enriched in metal contaminated MetalEurop sediments, suggesting their role in the long-term resilience of the community [27]. In the present study, IncP plasmid copy depletion in metal-treated microcosms may be explained by the cumulative costs of metal-stress metabolic response, plasmid maintenance, expression and conjugation [27].

Evidences of facilitator bacteria were found in groups of OTUs that responded similarly to the experimental conditions over time (TRGs). Metals were shown to prevent the development of opportunistic anaerobic, acetolactic methanogens Euryarchaeota (mainly the genus Methanotrix), known for their metal sensitivity [30] (TRG A). Such growth limitations were previously observed in-situ [7]. In addition, metals progressively favored the development of metal-resistant facilitator bacteria that precipitate metals or form large biofilms in the sedimentary structure (B TRGs). For example, in TRG B1 and 3, members of Zoogloea and

Dechloromonas were previously found to be highly represented in copper-contaminated activated sludges [31]. Zoogloea has been reported for its propensity to produce massive EPS matrices [32] and to accumulate metal ions [33]. Legionella is known for its ability to form biofilms [34], and Brevundimonas has been associated with lead removal in wastewater treatment plants [35]. Anaerolineaceae are known for cadmium removal [36]. Several SRBs such as Desulfosporinus, Desulfovibrio, and Desulfobulbus, are known to proliferate in the presence of metals because they produce hydrogen sulfide that may precipitate with metallic

101

ions [30]. Moreover, Clostridium seems to be a key player in this process by starting sulfate reduction and raising the pH of the environment in a context of metal contamination [37]. Most of the genera that were selected by metals (B TRGs), including those providing public goods, display species that were observed in wastewater treatment plants (WWTP) or in microbial communities in manure. For instance, Zoogloea, Dechloromonas, Terrimonas, and

Prosthecobacter were all found in activated sludge [31, 38–40], as well as Roseomonas [41,

42], Simplicispira [43], and Geothrix [44]. Zoogloea, Aquabacterium [45], Legionella,

Dechloromonas, Clostridium and Methyloversatilis have been found in WWTP [45–47]. Metals selected for exogenous bacteria arriving from upstream sediment and lands that potentially obtain their metal resistance through anthropologically impacted environments [48]. The resilience of the community would then have occurred through the early selection for metal- resistant bacteria, providing common protection for the rest of the microbial community such as EPS formation or metal detoxification allowing cooperation and reducing the metallic stress; rather than the use of broad-host-range IncP plasmids. This is supported by the Q-PCR results.

While the amount of the broadly distributed czcA gene was poorly affected by the experimental conditions, the pbrA gene, present within a narrower range of bacteria, was found to be enriched in metal conditions. In the TRG B1 and B4 groups, many Flavobacteriaceae and

Brevundimonas species were observed and were previously found to influence Pb motility [35,

49].

3.3. Microbial network: Linking bacteria acting in the public good with the rest of the community To assess the influence of bacteria acting as facilitator for the resilience of sedimentary microbiomes, we employed eLSA statistics. The network pinpointed Hyphomicrobiaceae

(Rhizobiales) as a hub OTU that preceded or co-occurred with many other members of the microbiome. In the metal-contaminated microcosms, this OTU was replaced by another

Rhizobiales, Xanthobacter of which several strains are known to have an increased capacity to

102

deal with metals [50]. The central position of a Hyphomicrobiaceae and Xanthobacter OTUs in the microbiome may be explained by their capacity to fix nitrogen, which may beneficiate the rest of the community [51], making them keystone OTUs for the community. The subdivision 3 of Verrucomicrobia, previously identified in fresh-water sediments [52], was also found to be a key player of both control and experimental communities. In metal-contaminated sediments, Xanthobacter, Verrucomicrobia subdivision 3 member and other unidentified hub

OTUs displayed many links with public-good providing bacteria such as Zoogloea,

Dechloromonas or SRBs. The latter facilitators may then be crucial for reducing the environmental pressure operated by metals for the keystone species of the community. In addition, Anaerolineaceae (TRG B3, Chloroflexi), which was previously shown to be a key component of metal contaminated communities and was suggested to impact metal solubility[36], displayed a positive local similarity index with all phyla represented in the network revealing its importance for the rest of the microbial community. Anaerolineaceae co- occurred with, Methanobacterium (OTU 266) and Methanolinea (OTU 533), which are hydrogenotrophic methanogens Euryarchaeota. This potential interaction may then be highly beneficial for this group that was negatively impacted by metals from the beginning of the contamination. Finally, the high degree of the OTU 169 (TRG B4), member of Veillonellaceae, with the rest of the community, highlighted the risk of dispersion of antibiotic resistance genes as suggested in-situ [7] that might be co-selected with metal-resistance genes [53]. Indeed,

Veillonellaceae are marker of fecal contamination and known to carry lots of antibiotic resistance genes [54], such as identified in Clostridia [10], or Anaerolineaceae found in WWTP

[39].

4. Conclusion The present microcosm experiment highlighted the presence of potential facilitator bacteria such as Zoogloea, Dechloromonas, Anaerolineaceae or SRBs and supported that the

103

coalescing of indigenous and upstream bacteria favors the resilience of metal-impacted sedimentary microbiomes, confirming in-situ observations. Facilitator bacteria then may play a key role by providing safe micro-niches for OTUs such as Rhizobiales and Verrucomicrobia, probable keystone organisms in the studied river-sediment microbiomes. Among them,

Anaerolineaceae were crucial for a large fraction the metal-stressed microbial community.

5. Materials and methods

5.1. Sampling methods and microcosm maintenance Sediments were sampled in the Sensée canal (Férin, France, 50°18’39.0’’N –

03°05’05.4’’E) on the fifth of December 2016 using a plexiglass core sampler (Ø = 7 cm) to avoid the mixing of sediment layers. Thirty sediment cores were collected, and the top first 1.5 cm were retrieved, homogenized and stored on ice during transport. Freshwater was sampled and stored at 4°C. On the sampling day, eight polypropylene microcosms (750 cm³) were filled with 120 mL of sediment and 500 mL of freshwater. Every day, 450 cm³ of the water column of each microcosm were replaced by water sampled in the river monthly, stored and oxygenized at 4°C with no sterilization step. During the first eight weeks of maintenance, a mixture of metals (ZnCl2, CuCl2(2H2O), PbCl2 and CdCl2(1/2H2O) was gradually added to four metal- treated microcosms at a rate of 20 mL every four days. Final metal concentrations are displayed in table S1. The pH remained stable over the course of the experiment (ca. pH 8.0). Microcosms were maintained in an air-conditioned room (23°C) over 6.5 months. The sediments were sampled for DNA extraction after two weeks (T0.5) and every month thereafter (T1.5, T2.5,

T3.5, T4.5, T5.5 and T6.5).

5.2. Total and bioavailable metal content in sediments For total metal quantification, 200 mg of dried and sieved (<63 μm) sediments were digested with concentrated hydrofluoric acid (Prolabo, Normapur) at 120°C for 48 h followed by aqua regia mineralization (hydrochloric + nitric acids, 2:1 v:v) at 120°C for 24 h. The total

104

metal concentrations (Al, Fe, Cd, Cu, Pb, and Zn) were determined using inductively coupled plasma atomic emission spectroscopy (ICP-AES; Varian, Vista Pro, axial view) according to the concentration range [55].

A proxy of the bioavailable metals (Al, Fe, Cd, Cu, Pb, and Zn) fraction in the sediments was determined by a soft acid attack: 1 g of dried sediment was mixed with 20 mL of HCl solution (1 M) and agitated for 24 h. The samples were filtered with a cellulose acetate membrane (0.45 μm). The metal concentrations were measured using ICP-AES [55] and are shown in Supplemental Table S1.

5.3. High-throughput 16S rRNA gene sequencing Sequencing libraries were prepared using a dual-PCR setup, targeting variable regions

V3 and V4 of the 16S rRNA gene, approx. 460bp. In the first step primers Uni341F (5’-

CCTAYGGGRBGCASCAG-3’) and Uni806R (5’-GGACTACNNGGGTATCTAAT-3’) originally published by Yu et al. [56] and modified as described in Sundberg et al. [57] were used. In a second PCR reaction step the primers additionally included Illumina specific sequencing adapters and a unique combination of index for each sample [7, 22]. First PCR reaction used Phusion® High-Fidelity DNA Polymerase (Thermo scientific, Carlsbad, CA,

USA) with an initial denaturation step at 98 °C for 30 s, followed by 35 amplification cycles consisting of denaturation at 98 °C for 10 s followed by 30 s at 61 °C and 15s at 72 °C and final extension of 7 min at 72 °C. Second PCR reaction used PCRBIO HiFi Polymerase (PCR

Biosystems Ltd, UK) with an initial denaturation step at 95 °C for 1 min, followed by 15 amplification cycles consisting of denaturation at 95 °C for 15 s followed by 15 s at 56 °C and

30 at 72 °. After each PCR reaction, amplicon products less than 200 bp were purified using

HighPrep™ PCR Clean-up System magnetic beads (MagBio, MD, USA) according to the manufacturer’s instructions. The samples were adjusted to equimolar concentrations using

SequalPrep Normalization Plate 96-well kit (Thermo Fisher Scientific). A final library pool was

105

made using 5µL of each individual normalized sample library, concentrated using DNA Clean

TM and Concentrator 5 kit (Zymo Research, Irvine, CA, USA) in a final volume of 10µL.

Finally, the pooled libraries were adjusted to 4nM and sequenced using MiSeq v2 reagents kit

(500 cycles) using 2 × 250 bp paired-end mode on an Illumina MiSeq platform (Illumina, San

Diego, CA, USA) according to the manufacturer’s instructions. The amplicon sequences were analyzed with the QIIME pipeline (https://github.com/maasha/qiime pipe) as previously described [22]. The MiSeq Controller Software was used to perform the sequence demultiplexing and sequencing adapters and primers were trimmed using cutadapt v1.18 [58] and any pairs were any of two primers couldn’t be found were discarded. Trimmed reads were then processed using a custom BioDSL pipeline (http://maasha.github.io/BioDSL/).

Specifically, paired-end reads were merged using merge_pair_seq, assembled pairs with a total length shorter than 300 bp or an average Phred score quality lower than Q25 were discarded.

Clean reads were clustered in OTU using a 97% sequence similarity threshold using cluster_otus and USEARCH v7.0.1090 [59] and chimeric OTUs removed using UCHIME [60].

Each OTU cluster was annotated using mothur classify.seq() [61], an implementation of the

RDP Classifier (method=wang) [62] against the Ribosomal Database Project Trainset 16 database [63]. Representative OTU sequence were aligned against a reference alignment using mothur and an approximate maximum likelihood phylogenetic tree was built using FastTree

[64]. Samples (4 biological replicates x 2 technical replicates) with less than 3 000 OTU read counts were not included, as they do not provide enough coverage for further diversity analysis

[65]. Information regarding the sequence counts for each sample is shown in Supporting Table

S2.

106

5.4. Measuring the diversity through time Alpha-diversity estimates based on OTUs richness and the Shannon’s diversity index were assessed using a rarefied dataset (R-package “Vegan” in RGui software [66]) containing

5150 sequences using the Past3 software [67]. The beta-diversity (a measure of species turnover over time) was estimated. For this, a PERMANOVA test was performed using the R-package

“Vegan” with a log10(1+x)-transformed database (as OTU abundance variation was higher than

1 000-fold) with 10 000 permutations (Table S3). Differences in phyla between the control and metal-treated microcosms were assessed according to an FDR-corrected p-value inferred from a t-test (R-package “Vegan”), and the different classes of Proteobacteria as well as the new order of β-Protobacteria were considered separately (Table S4). The evolution of phylogenetic relatedness in the control and metal-treated microcosms was then assessed between the in-situ profile and at each time point for both conditions using the phylogenetic tree obtained from

OTU cluster representative sequence reads with the R-package “ape”. MPD quantifies the mean relatedness in a group of OTUs by considering all possible pairs of OTUs. MNTD quantifies the mean relatedness by considering only the nearest phylogenetic neighbors We calculated the weighted net relatedness index (NRI) and the nearest taxon index (NTI), which assess the number of standard deviations when comparing the mean pairwise distance (MPD) or the mean nearest taxon distance (MNTD), respectively, to the corresponding null distribution with 1 000 randomizations, 95% confidence interval, p < 0.05 [68, 69] for each time point. An NRI or NTI value >2 indicates that the coexisting taxa are significantly closely related, while an NRI or NTI

<-2 indicates significant phylogenetic overdispersion [70]. The metrics were calculated with the R-package “Picante” [71]. The between-community version of NRI (βNRI) and NTI

(βNTI), βNTI [70], were determined (weighted, with 1 000 randomizations, R-package

“MicEco” [72] ) comparing each time point to time 0. A βNTI value >2 indicates a greater dispersion than expected by chance, while a βNTI <-2 indicates a static community.

Intermediate values indicate that changes appear to be due to stochastic events [70].

107

5.5. Identification and validation of the time response groups The 16S rRNA gene amplicon dataset was divided by time point, and the OTU profiles obtained after 0.5, 2.5, 3.5 and 6.5 months were assessed separately. The OTUs that responded significantly to metal stress were assessed using a negative binomial distribution and a generalized linear model, which were corrected with 1 000 resampling iterations of the residual variance (nbGLM, p < 0.05) [7]. The 600 selected OTUs were plotted in a generalized heatmap, and the Treatment Response Groups (TRG) were defined with a hierarchical cluster dendrogram (Euclidian distance and average clustering, R-package “gplots” [73]). The statistical validity of the obtained TRGs was tested against the null model by a Monte Carlo simulation including all OTUs [7, 22].

5.6. Dynamic occurrence network The 16S rRNA gene amplicon dataset was then used to infer potential interactions between OTUs using an Extended Local Similarity Analysis (eLSA) [74]. The eLSA approach was applied using 1 technical replicate by biological replicate, to the most abundant OTUs in the 16S rRNA library with the eLSA implementation as previously described [75], using the following parameters: “lsa_compute data.txt result.txt -s 8 -r4 -p perm -x 1000 -d 4” [76–78].

The minimum LSA index was set at 0.8, and the p-value threshold was set at 0.05. The network visualization and analysis was performed with Cytoscape 3.6.1 [79] using an organic-type layout and analyzed with the NetworkAnalyzer tool.

5.7. Functional and plasmid-associated gene content assessment by quantitative PCR. The abundance of the metal resistance genes czcA and pbrA and the origin of transfer

(oriT) of broad host range IncP plasmids were assessed by quantitative PCR. The primers used for the PCR amplification of czcA and oriT were retrieved from the literature (Roosa et al.,

2014, Götz et al., 1996) (Table S5). The primers for the PCR amplification of pbrA were designed for this study. For pbrA, the genetic sequences of Cupriavidus metallidurans CH34,

108

Klebsiella pneumonia and Enterobacter cloacae were aligned using the ClustalX. The best consensus sequences were used as templates for the design of degenerated primers. The physicochemical properties of each pair of primers were compared using the ThermoFisher

Scientific Multiple Primer Analyzer tool [80]. The best candidates were selected (Table S5).

The specificity of the primers was tested using BLAST (NCBI) and classical PCR of pure

Cupriavidus metallidurans CH34 and Pseudomonas putida KT2440 cultures.

For quantitative PCR (Q-PCR), the DNA samples were purified using a QIAquick®

PCR Purification (QIAGEN) kit to eliminate potential PCR inhibitors. A StepOnePlus Real-

Time PCR System (Thermo Fisher Scientific) was used for the quantification of oriT in IncP, czcA and pbrA. The relative concentrations of these genes were calculated as the ratio of the expression of the target gene and that of the reference 16S rRNA gene using the universal primers 518R (5’-ATTACCGCGGCTGCTGG-3’) and 341F (5’- CCTACGGGAGGCAGCAG

-3’). The results are presented as a double ratio of the expression of the target gene (IncP oriT, czcA or pbrA) and the reference gene (16S rRNA) at a given time in comparison with the expression at T0. The ratios were calculated according to equation (1) [81]

ΔCP target (T0 – Tx) ΔCP ref (T0 – Tx) Ratio = (Etarget) / (Eref) (1)

where E is the PCR efficiency calculated according to Ramakers et al., 2003, and CP is the point at which the amplification curve crosses the threshold.

Duplicate PCRs were performed for each DNA sample and target gene combination using SYBRTM Green master mix (Applied BiosystemsTM). Each replicate consisted of serial dilutions (0.5, 0.25, 0.125 and 0.0625 ng/µL for oriT and pbrA and 1, 0.7, 0.5, 0.25 ng/µL for czcA). After an initial denaturation step at 94°C for 10 min, 40 cycles of 15 seconds at 95°C,

30 seconds at 60°C and 30 seconds at 72°C were performed followed by a final denaturation step (60 to 90°C, +0.3°C/min).

109

List of Abbreviations eLSA: extended Local Similarity Analysis ; ICP-AES: inductively coupled plasma atomic emission spectroscopy ; HGT: Horizontal Gene Transfer ; nbGLM: negative binomial

Generalized Linear Model ; NRI: Net Relatedness Index ; NTI: Nearest Taxon Index; SRB:

Sulfate Reducing Bacteria ; TRG: Treatment Response Group; WWTP: WasteWater Treatment

Plant.

Declarations

Availability of data and material Supplemental information is available as Supporting Table file and Supporting Figure file.

Unassembled raw amplicon data were deposited at the Sequence Read Archive public repository (SRA, https://www.ncbi.nlm.nih.gov/sra) under the accession number

PRJNA574447 (https://www.ncbi.nlm.nih.gov/sra/PRJNA574447).

Competing interests The authors declare no conflicts of interest.

Funding This research was funded by the Fund for Scientific Research (F.R.S-FNRS) FRFC

7050357, PF014F828/A09016F and T.0127.14 (Belgium) and the Pole d’ Attraction

Interuniversitaire (PAI) n◦ P7/25.

Authors' contributions V.C., A.G., R.W. designed research; V.C., A.G., G.B., J.N. performed research; V.C., A.G.,

G.B., J.N. analyzed data; and V.C., A.G., G.B., D.C.G., J.N., S.J.S., R.W. wrote the paper.

Acknowledgements Authors wish to express a deep gratitude to Nicolas Decruyenaere for its help with computing.

110

References

1. Gillan DC, Roosa S, Kunath B, Billon G, Wattiez R. The long-term adaptation of bacterial communities in metal-contaminated sediments: A metaproteogenomic study. Environ Microbiol 2015; 17: 1991–2005. 2. Araya R, Tani K, Takagi T, Yamaguchi N, Nasu M. Bacterial activity and community composition in stream water and biofilm from an urban river determined by fuorescent in situ hybridization and DGGE analysis. FEMS Microbiol Ecol 2003; 43: 111–119. 3. Besemer K. Aquatic Biofilms: Ecology, Water Quality and Wastewater Treatment. Aquatic Biofilms: Ecology, Water Quality and Wastewater Treatment . 2016. Caister Academic Press, Norfolk, UK. 4. Brown BL, Swan CM, Auerbach DA, Campbell Grant EH, Hitt NP, Maloney KO, et al. Metacommunity theory as a multispecies, multiscale framework for studying the influence of river network structure on riverine communities and ecosystems. J North Am Benthol Soc 2011; 30: 310–327. 5. De Oliveira LFV, Margis R. The source of the river as a nursery for microbial diversity. PLoS One 2015; 10: 1–11. 6. Mansour I, Heppell CM, Ryo M, Rillig MC. Application of the Microbial Community Coalescence Concept to Riverine Networks. Biol Rev 2018; 93: 1832–1845. 7. Jacquiod S, Cyriaque V, Riber L, Al-soud WA, Gillan DC, Wattiez R, et al. Long-term industrial metal contamination unexpectedly shaped diversity and activity response of sediment microbiome. J Hazard Mater 2018; 344: 299–307. 8. Wang L, Zhang J, Li H, Yang H, Peng C, Peng Z, et al. Shift in the microbial community composition of surface water and sediment along an urban river. Sci Total Environ 2018; 627: 600–612. 9. Marti E, Jofre J, Balcazar JL. Prevalence of Antibiotic Resistance Genes and Bacterial Community Composition in a River Influenced by a Wastewater Treatment Plant. PLoS One 2013; 8: 1–8. 10. Jacquiod S, Brejnrod A, Morberg SM, Abu Al-Soud W, Sørensen SJ, Riber L. Deciphering conjugative plasmid permissiveness in wastewater microbiomes. Mol Ecol 2017; 26: 3556– 3571. 11. Sherameti I. Heavy Metal Contamination of Soils. 2015. Springer International Publishing. 12. Oves M, Mabood Hussain F. Antibiotics and Heavy Metal Resistance Emergence in Water Borne Bacteria. J Investig Genomics 2016; 3: 3–5. 13. Stepanauskas R, Glenn TC, Jagoe CH, Tuckfield RC, Lindell AH, King CJ, et al. Coselection for

111

microbial resistance to metals and antibiotics in freshwater microcosms. Environ Microbiol 2006; 8: 1510–1514. 14. Gillan DC. Metal resistance systems in cultivated bacteria: Are they found in complex communities? Curr Opin Biotechnol 2016; 38: 123–130. 15. Baker-Austin C, Wright MS, Stepanauskas R, McArthur J V. Co-selection of antibiotic and metal resistance. Trends Microbiol 2006; 14: 176–182. 16. Martinez JL. The role of natural environments in the evolution of resistance traits in pathogenic bacteria. Proc R Soc B Biol Sci 2009; 276: 2521–2530. 17. Seiler C, Berendonk TU. Heavy metal driven co-selection of antibiotic resistance in soil and water bodies impacted by agriculture and aquaculture. Front Microbiol 2012; 3: 1–10. 18. Perry JA, Wright GD. The antibiotic resistance ‘mobilome’: Searching for the link between environment and clinic. Front Microbiol 2013; 4: 1–7. 19. O’Brien S, Hesse E, Luján A, Hodgson DJ, Gardner A, Buckling A. No effect of intraspecific relatedness on public goods cooperation in a complex community. Evolution (N Y) 2018; 72: 1165–1173. 20. Smith P, Schuster M. Public goods and cheating in microbes. Curr Biol 2019; 29: R442–R447. 21. Berg J, Brandt KK, Al-Soud WA, Holm PE, Hansen LH, Sørensen SJ, et al. Selection for Cu- tolerant bacterial communities with altered composition, but unaltered richness, via long- term cu exposure. Appl Environ Microbiol 2012; 78: 7438–7446. 22. Nunes I, Jacquiod S, Brejnrod A, Holm PE, Johansen A, Brandt KK, et al. Coping with copper: Legacy effect of copper on potential activity of soil bacteria following a century of exposure. FEMS Microbiol Ecol 2016; 92: 1–12. 23. Rillig MC, Antonovics J, Caruso T, Lehmann A, Powell JR, Veresoglou SD, et al. Interchange of entire communities: Microbial community coalescence. Trends Ecol Evol 2015; 30: 470–476. 24. Smalla K, Haines AS, Jones K, Krögerrecklenfort E, Heuer H, Schloter M, et al. Increased abundance of IncP-1β plasmids and mercury resistance genes in mercury-polluted river sediments: First discovery of IncP-1β plasmids with a complex mer transposon as the sole accessory element. Appl Environ Microbiol 2006; 72: 7253–7259. 25. Grohmann E. Horizontal Gene Transfer Between Bacteria Under Natural Conditions. In: Ahmad I, Ahmad F, Pichtel J (eds). Microbes and Microbial Technology: Agricultural and Environmental Applications. 2011. pp 163–187. 26. Swenson NG. Phylogenetic resolution and quantifying the phylogenetic diversity and dispersion of communities. PLoS One 2009; 4: e4390. 27. Cyriaque V, Jacquiod S, Riber L, Abu Al-soud W, Gillan DC, Sørensen SJ, et al. Selection and propagation of IncP conjugative plasmids following long-term anthropogenic metal pollution

112

in river sediments. J Hazard Mater 2020; 382: 121173. 28. Popowska M, Krawczyk-Balska A. Broad-host-range IncP-1 plasmids and their resistance potential. Front Microbiol 2013; 4: 1–8. 29. Jechalke S, Dealtry S, Smalla K, Heuer H. Quantification of IncP-1 plasmid prevalence in environmental: Samples. Appl Environ Microbiol 2013; 79: 1410–1413. 30. Paulo LM, Stams AJM, Sousa DZ. Methanogens, sulphate and heavy metals: a complex system. Rev Environ Sci Biotechnol 2015; 14: 537–553. 31. Ouyang F, Ji M, Zhai H, Dong Z, Ye L. Dynamics of the diversity and structure of the overall and nitrifying microbial community in activated sludge along gradient copper exposures. Appl Microbiol Biotechnol 2016; 100: 6881–6892. 32. Dugan P. The Genus Zoogloea. In: Stoner D, Pickrum H (eds). The Prokaryotes. 1981. pp 960– 970. 33. Norberg AB, Persson H. Accumulation of heavy‐metal ions by Zoogloea ramigera. Biotechnol Bioeng 1984; 26: 239–246. 34. Mampel J, Spirig T, Weber SS, Janus AJ, Molin S, Hilbi H. Planktonic Replication Is Essential for Biofilm Formation by Legionella pneumophila in a Complex Medium under Static and Dynamic Flow Conditions Planktonic Replication Is Essential for Biofilm Formation by Legionella pneumophila in a Complex Medium under. Society 2006; 72: 2885–2895. 35. Resmi G, Thampi SG, Chandrakaran S. Brevundimonas vesicularis:A novel bio-sorbent for removal of lead from wastewater. Int J Environ Res 2010; 4: 281–288. 36. Meng D, Li J, Liu T, Liu Y, Yan M, Hu J, et al. Effects of redox potential on soil cadmium solubility: Insight into microbial community. J Environ Sci (China) 2018; 75: 224–232. 37. Alexandrino M, Costa R, Canário AVM, Costa MC. Clostridia initiate heavy metal bioremoval in mixed sulfidogenic cultures. Environ Sci Technol 2014; 48: 3378–3385. 38. Zhang T, Shao MF, Ye L. 454 Pyrosequencing reveals bacterial diversity of activated sludge from 14 sewage treatment plants. ISME J 2012; 6: 1137–1147. 39. Wang X, Hu M, Xia Y, Wen X, Ding K. Pyrosequencing analysis of bacterial diversity in 14 wastewater treatment systems in china. Appl Environ Microbiol 2012; 78: 7042–7047. 40. Lee J, Park B, Woo SG, Lee J, Park J. Prosthecobacter algae sp. nov., isolated from activated sludge using algal metabolites. Int J Syst Evol Microbiol 2014; 64: 663–667. 41. Chu CW, Chen Q, Wang CH, Wang HM, Sun ZG, He Q, et al. Roseomonas chloroacetimidivorans sp. nov., a chloroacetamide herbicide-degrading bacterium isolated from activated sludge. Antonie van Leeuwenhoek, Int J Gen Mol Microbiol 2016; 109: 611– 618. 42. Jiang X, Ma M, Li J, Lu A, Zhong Z. Bacterial Diversity of Active Sludge in Wastewater

113

Treatment Plant. Earth Sci Front 2008; 15: 163–168. 43. Lu S, Ryu SH, Chung BS, Chung YR, Park W, Jeon CO. Simplicispira limi sp. nov., isolated from activated sludge. Int J Syst Evol Microbiol 2007; 57: 31–34. 44. Paiva MC, Ávila MP, Reis MP, Costa PS, Nardi RMD, Nascimento AMA. The microbiota and abundance of the class 1 integron-integrase gene in tropical sewage treatment plant influent and activated sludge. PLoS One 2015; 10: 1–12. 45. Cydzik-Kwiatkowska A, Zielińska M. Bacterial communities in full-scale wastewater treatment systems. World J Microbiol Biotechnol 2016; 32: 1–8. 46. Jacquiod S, Brejnrod A, Morberg SM, Al-Soud WA, Sorensen SJ, Riber L. Deciphering conjugative plasmid permissiveness dynamics in wastewater microbiomes. Mol Ecol 2017; 1– 16. 47. Morico K, Moor B, Novotny J. Electronics and Metal Finishing and Processing. Water Environ Res 2009; 81: 1642–1653. 48. Gupta SK, Shin H, Han D, Hur H-G, Unno T. Metagenomic analysis reveals the prevalence and persistence of antibiotic- and heavy metal-resistance genes in wastewater treatment. J Microbiol 2018; 56: 408–415. 49. Fonti V, Beolchini F, Rocchetti L, Dell’Anno A. Bioremediation of contaminated marine sediments can enhance metal mobility due to changes of bacterial diversity. Water Res 2015; 68: 637–650. 50. Cubillas C, Miranda-Sánchez F, González-Sánchez A, Elizalde JP, Vinuesa P, Brom S, et al. A comprehensive phylogenetic analysis of copper transporting P1BATPases from bacteria of the Rhizobiales order uncovers multiplicity, diversity and novel taxonomic subtypes. Microbiologyopen 2017; 6: 1–13. 51. Teng Y, Wang X, Li L, Li Z, Luo Y. Rhizobia and their bio-partners as novel drivers for functional remediation in contaminated soils. Front Plant Sci 2015; 6: 1–11. 52. He S, Stevens SLR, Chan L-K, Bertilsson S, Glavina del Rio T, Tringe SG, et al. Ecophysiology of Freshwater Verrucomicrobia Inferred from. mSphere 2017; 2: e00277-17. 53. Pal C, Asiani K, Arya S, Rensing C, Stekel DJ, Larsson DGJ, et al. Metal Resistance and Its Association With Antibiotic Resistance, 1st ed. Advances in Microbial Physiology . 2017. Elsevier Ltd. 54. Marchandin H, Jumas-Bilak E. Family Veillonellaceae. In: Rosenberg E, DeLong EF, Lory S, Stackebrandt E, Thompson F (eds). The Prokaryotes: Firmicutes and Tenericutes. 2014. Springer-Verlag, Berlin, pp 1–567. 55. Roosa S, Prygiel E, Lesven L, Wattiez R, Gillan D, Ferrari BJD, et al. On the bioavailability of trace metals in surface sediments: a combined geochemical and biological approach. Environ

114

Sci Pollut Res 2016; 23: 10679–10692. 56. Yu Y, Lee C, Kim J, Hwang S. Group-specific primer and probe sets to detect methanogenic communities using quantitative real-time polymerase chain reaction. Biotechnol Bioeng 2005; 89: 670–679. 57. Sundberg C, Al-Soud WA, Larsson M, Alm E, Yekta SS, Svensson BH, et al. 454 Pyrosequencing Analyses of Bacterial and Archaeal Richness in 21 Full-Scale Biogas Digesters. FEMS Microbiol Ecol 2013; 85: 612–626. 58. Martin M. The relationship between organizational culture and knowledge management,& their simultaneous effects on customer relation management. Adv Environ Biol 2013; 7: 2803– 2809. 59. Edgar RC. Search and clustering orders of magnitude faster than BLAST. Bioinformatics 2010; 26: 2460–2461. 60. Edgar RC, Haas BJ, Clemente JC, Quince C, Knight R. UCHIME improves sensitivity and speed of chimera detection. Bioinformatics 2011; 27: 2194–2200. 61. Schloss PD, Westcott SL, Ryabin T, Hall JR, Hartmann M, Hollister EB, et al. Introducing mothur: Open-source, platform-independent, community-supported software for describing and comparing microbial communities. Appl Environ Microbiol 2009; 75: 7537–7541. 62. Wang Q, Garrity GM, Tiedje JM, Cole JR. Naïve Bayesian classifier for rapid assignment of rRNA sequences into the new bacterial taxonomy. Appl Environ Microbiol 2007; 73: 5261– 5267. 63. Cole JR, Wang Q, Fish JA, Chai B, McGarrell DM, Sun Y, et al. Ribosomal Database Project: Data and tools for high throughput rRNA analysis. Nucleic Acids Res 2014; 42: 633–642. 64. Price MN, Dehal PS, Arkin AP. FastTree 2 - Approximately maximum-likelihood trees for large alignments. PLoS One 2010; 5: e9490. 65. Caporaso JG, Lauber CL, Walters WA, Berg-Lyons D, Lozupone CA, Turnbaugh PJ, et al. Global patterns of 16S rRNA diversity at a depth of millions of sequences per sample. Proc Natl Acad Sci 2011; 108: 4516–4522. 66. Oksanen J, Blanchet FG, Friendly M, Kindt R, Legendre P, McGlinn D, et al. vegan: Community Ecology Package. R package version 2.5-6. 2019. 67. Hammer Ø, Harper DAT a. T, Ryan PD. PAST: Paleontological Statistics Software Package for Education and Data Analysis. Palaeontol Electron 2001; 4(1): 1–9. 68. Stegen JC, Lin X, Fredrickson JK, Konopka AE. Estimating and mapping ecological processes influencing microbial community assembly. Front Microbiol 2015; 6: 1–15. 69. Webb CO, Ackerly DD, Kembel SW. Phylocom: Software for the analysis of phylogenetic community structure and trait evolution. Bioinformatics 2008; 24: 2098–2100.

115

70. Stegen JC, Lin X, Konopka AE, Fredrickson JK. Stochastic and deterministic assembly processes in subsurface microbial communities. ISME J 2012; 6: 1653–1664. 71. Kembel SW, Cowan PD, Helmus MR, Cornwell WK, Morlon H, Ackerly DD, et al. Picante: R tools for integrating phylogenies and ecology. Bioinformatics 2010; 26: 1463–1464. 72. Russel J. Russel88/MicEco v0.9.2. 2018. 73. Warnes GR, Bolker B, Bonebakker L, Gentleman R, Liaw WHA, Lumley T, et al. Package ‘gplots’: Various R programming tools for plotting data. R Packag version 2170 2016; 1–68. 74. Xia LC, Ai D, Cram J, Fuhrman JA, Sun F. Efficient statistical significance approximation for local similarity analysis of high-throughput time series data. Bioinformatics 2013; 29: 230–237. 75. Wang H, Wei Z, Mei L, Gu J, Yin S, Faust K, et al. Combined use of network inference tools identifies ecologically meaningful bacterial associations in a paddy soil. Soil Biol Biochem 2017; 105: 227–235. 76. Durno WE, Hanson NW, Konwar KM, Hallam SJ. Expanding the boundaries of local similarity analysis. BMC Genomics 2013; 14: 1–14. 77. Ruan Q, Dutta D, Schwalbach MS, Steele JA, Fuhrman JA, Sun F. Local similarity analysis reveals unique associations among marine bacterioplankton species and environmental factors. Bioinformatics 2006; 22: 2532–2538. 78. Eiler A, Heinrich F, Bertilsson S. Coherent dynamics and association networks among lake bacterioplankton taxa. ISME J 2012; 6: 330–342. 79. Shannon P, Markiel A, Owen Ozier 2, Baliga NS, Wang JT, Ramage D, et al. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res 2003; 2498–2504. 80. Breslauer KJ, Frank R, Blocker H, Marky LA. Predicting DNA duplex stability from the base sequence. Proc Natl Acad Sci 1986; 83: 3746–3750. 81. Pfaffl MW. A new mathematical model for relative quantification in real-time RT–PCR. Nucleic Acids Res 2001; 29: 16–21.

116

Supporting Figure File Valentine Cyriaque, Augustin Géron, Gabriel Billon, David C. Gillan, Joseph Nesme, Søren J. Sørensen and Ruddy Wattiez. “Metal-induced bacterial interactions promote diversity in river-sediment microbiomes”.

117

Figure S1: Generalized heatmap of significantly responding OTUs across experimental design. Significance of OTU response patterns were extracted by means of generalized linear models under negative binomial distribution (nbGLM, p < 0.05, 1,000 iterations) applied by comparing control and metal-added microcosm after 0.5, 3.5 and 6.5 months. Time response grouping was done with Euclidean distance and average clustering using based on centre-scaled taxa abundance. Two main Time Response Groups (TRG) have been defined depending on the maintenance or enrichment of OTUs in control microcosms (Group A) or the maintenance of OTUs in metal-added microcosms (Group B) directly selected (B1), intermittently selected (Group B2), progressively selected (B3) and lately selected (B4).

118

Monte-Carlo Simulation p-value= 9.999e-05

TRG A TRG B1 TRG B2

TRG B3 TRG B4

Others F3=10.82 % F3=10.82

F2=36.42 %

F1=51.20 %

Figure S2: Statistical verification and validation of the abundance driven pattern identified with the five TRGs by means of constrained BCA followed by a random null model Monte-Carlo simulation implementing 100,000 group permutations (p < 1.0E-05).

119

Figure S3: Abundances (read counts) composing TRGs and their corresponding taxonomical orders (for Proteobacteria, class is taken into consideration). OTUs significantly responding to metal contamination over time were highlighted using a nbGLM (p<0.05) and TRGs were identified by displaying those OTUs in a heatmap (Supplemental figure S2).

120

OTU 207 OTU

Anaerolineaceae

϶ Chloroflexi

Acidobacteria Actinobacteria Alpha-proteobacteria Archae Bacteria Bacteroidetes Beta-proteobacteria Candidatus-Saccharibacteria Chloroflexi Crenarchaeota Cyanobacteria Delta-proteobacteria Euryarchaeota Firmicutes Gamma-proteobacteria Ignavibacteriae Latescibacteria Nitrospira Parcubacteria Spirochaetes Thaumarchaeota Verrucomicrobia

OTU 46 OTU

Verrucomicrobia

Thaumarchaeota

Spirochaetes

Parcubacteria

Nitrospira

Latescibacteria

Ignavibacteriae

Gamma-proteobacteria

Firmicutes

Euryarchaeota

Delta-proteobacteria

Cyanobacteria

Crenarchaeota

Chloroflexi

Candidatus-Saccharibacteria

Beta-proteobacteria

Bacteroidetes

Bacteria

Archae

Alpha-proteobacteria

Actinobacteria

Acidobacteria

Node taxonomy Node

y y

. Each Each .

339

) between

OTU OTU

and and red line

OTU 90 & 131 & 90 OTU

with with a time dela

value value < 0.0

mental microcosms (B) microcosms mental

a a local similarity

represents represents

ode size only permits to get attention. toget permits only size ode

line

local similarity local scoressimilarity (>0.8, p

represent represent positive association of OTU

irected irected

lines

based based on

Green

for the control microcosms (A) and experi and (A) microcosms control the for

OTU.

varying varying graph

nts nts a

egative egative associations.

: :

most represented OTUs represented most Figure node repre represent n n Increased OTUs. of succession representing A

121

Figure S5: α-diversity obtained after 6-months incubation (16°C) of Férin sediment-filled microcosm as preliminary experiment. For this work, water renewal was performed with autoclave sterilized water obtained from the Sensée Canal (Férin, France). 16S rRNA amplicon sequencing and α-diversity analysis on rarefied data (11 500 counts) have been performed as described in the Material and Method section. p-value<0.05.

122

Supporting Table File

Valentine Cyriaque, Augustin Géron, Gabriel Billon, David C. Gillan, Joseph Nesme, Søren J. Sørensen and Ruddy Wattiez. “Metal-induced bacterial interactions promote diversity in river-sediment microbiomes”.

Table S1: Total and bioavailable (HCl extracted) metal concentration (mg/kg (±SEM)) quantified in different microcosms before and after incubation time using ICP-AES. mg/kg (±SEM) Cd Cu Pb Zn In-situ 1.480 ±0.030 17.549 ±0.454 123.743 ±2.515 478.783 ±11.865 Total T6-Control 1.526 ±0.039 17.610 ±0.813 125.800 ±1.320 512.872 ±8.383 143.318 327.917 3076.306 6124.106 T6 metal ±3.110 ±9.491 ±85.089 ±76.149 In-situ 0.433 ±0.022 10.216 ±0.167 90.318 ±1.617 160.849 ±1.569 Bioavailable T6-Control 0.383 ±0.009 10.371 ±0.046 90.162 ±0.923 166.161 ±2.311 124.671 281.113 2625.419 4906.686 T6 metal ±1.972 ±6.316 ±58.180 ±41.453

123

Table S2: Summary table of 16S rRNA gene amplicon profiles obtained

Condition Time (month) Read Count Suitability Condition Time (month) Count Suitability In-situ 0 11298 Yes In-situ 0 7651 Yes In-situ 0 12131 Yes In-situ 0 11379 Yes Control 0.5 10059 Yes Metal 0.5 16496 Yes Control 0.5 8683 Yes Metal 0.5 11640 Yes Control 0.5 No Metal 0.5 4528 Yes Control 0.5 No Metal 0.5 4886 Yes Control 0.5 5701 Yes Metal 0.5 5601 Yes Control 0.5 4036 Yes Metal 0.5 5776 Yes Control 0.5 6966 Yes Metal 0.5 20201 Yes Control 0.5 3546 Yes Metal 0.5 25809 Yes Control 1.5 23089 Yes Metal 1.5 28104 Yes Control 1.5 25302 Yes Metal 1.5 9679 Yes Control 1.5 No Metal 1.5 9800 Yes Control 1.5 No Metal 1.5 10658 Yes Control 1.5 16782 Yes Metal 1.5 16678 Yes Control 1.5 12349 Yes Metal 1.5 14162 Yes Control 1.5 16424 Yes Metal 1.5 54439 Yes Control 1.5 9828 Yes Metal 1.5 43290 Yes Control 2.5 13923 Yes Metal 2.5 12041 Yes Control 2.5 3614 Yes Metal 2.5 10633 Yes Control 2.5 2226 Yes Metal 2.5 5267 Yes Control 2.5 11376 Yes Metal 2.5 5837 Yes Control 2.5 12989 Yes Metal 2.5 13587 Yes Control 2.5 8338 Yes Metal 2.5 5921 Yes Control 2.5 10708 Yes Metal 2.5 36494 Yes Control 2.5 14063 Yes Metal 2.5 23994 Yes Control 3.5 59483 Yes Metal 3.5 33318 Yes Control 3.5 49262 Yes Metal 3.5 35682 Yes Control 3.5 53036 Yes Metal 3.5 44278 Yes Control 3.5 51477 Yes Metal 3.5 36406 Yes Control 3.5 35589 Yes Metal 3.5 40158 Yes Control 3.5 46354 Yes Metal 3.5 46302 Yes Control 3.5 24545 Yes Metal 3.5 19518 Yes Control 3.5 22963 Yes Metal 3.5 19078 Yes Control 4.5 39237 Yes Metal 4.5 30899 Yes Control 4.5 43816 Yes Metal 4.5 30057 Yes Control 4.5 49819 Yes Metal 4.5 50986 Yes Control 4.5 40198 Yes Metal 4.5 34363 Yes Control 4.5 44822 Yes Metal 4.5 47768 Yes Control 4.5 40940 Yes Metal 4.5 39588 Yes Control 4.5 21161 Yes Metal 4.5 18369 Yes Control 4.5 19978 Yes Metal 4.5 17053 Yes Control 5.5 48152 Yes Metal 5.5 26571 Yes Control 5.5 46540 Yes Metal 5.5 20045 Yes Control 5.5 50735 Yes Metal 5.5 41689 Yes Control 5.5 45433 Yes Metal 5.5 42978 Yes Control 5.5 42661 Yes Metal 5.5 39575 Yes Control 5.5 36708 Yes Metal 5.5 43782 Yes Control 5.5 14803 Yes Metal 5.5 8279 Yes Control 5.5 22621 Yes Metal 5.5 21002 Yes Control 6.5 46429 Yes Metal 6.5 31007 Yes Control 6.5 36269 Yes Metal 6.5 23598 Yes Control 6.5 51928 Yes Metal 6.5 42664 Yes Control 6.5 23107 Yes Metal 6.5 35562 Yes Control 6.5 40279 Yes Metal 6.5 35106 Yes Control 6.5 26767 Yes Metal 6.5 33755 Yes Control 6.5 20914 Yes Metal 6.5 16176 Yes Control 6.5 1518 No Metal 6.5 17668 Yes

124

Table S3: Dissimilarities between samples (Permanova using Bray-Curtis dissimilarity profiles and 105 permutations) evolution through time obtained from 16S rRNA amplicon sequencing carried out from DNA extracted from Férin in-situ sediment on the day of sampling (starting point), from control and metal-added microcosm sediments.

F.Model R2 Pr(>F)

Time 10.767 0.31760 1,00E-05 ***

Treatment 38.289 0.16134 1,00E-05 ***

Time:Treatment 4.776 0.12075 1,00E-05 ***

125

Table S4: Taxonomic composition of the prokaryotic community of sediments between control and metal-added microcosms. The table shows average relative abundance (±SE) distribution of taxonomic groups. Statistical significance was inferred by ANOVA followed by post-hoc testing and false

discovery rate adjusted p-value for multiple comparisons (FDR, p < 0.05).

4,70E-01

3,42E-01

1,93E-01

9,36E-03

4,81E-03

6,17E-02

2,05E-04

2,78E-01

5,62E-02

3,70E-02

4,22E-02

9,09E-04

2,47E-01

9,21E-03

2,52E-01

8,79E-01

7,44E-03

1,15E-02

3,70E-02

8,26E-02

8,01E-02

2,30E-04

8,26E-02

1,89E-02

3,95E-02

4,46E-02

2,31E-01

1,39E-02

3,70E-02

3,95E-02

2,31E-01

9,09E-04

2,05E-04

7,44E-03

6,52E-03

7,61E-01

5,15E-02

6,51E-01

5,15E-02

3,05E-01

6,52E-03

2,88E-01

1,39E-02

1,77E-02

8,81E-01

8,79E-01

6.6

7,42E-01

2,61E-01

4,55E-01

4,44E-02

9,59E-03

8,20E-04

8,83E-02

1,49E-02

7,91E-02

4,55E-01

3,02E-02

9,10E-01

9,59E-03

4,74E-01

7,49E-01

9,59E-03

3,81E-01

5,28E-01

3,02E-02

7,69E-02

1,26E-02

6,95E-02

3,91E-02

9,59E-03

4,74E-01

6,29E-01

2,98E-02

4,56E-01

1,56E-01

1,91E-02

3,02E-02

8,49E-03

6,87E-03

5,28E-01

3,02E-02

5,36E-01

1,11E-01

6,96E-01

1,25E-01

6,29E-01

6,73E-01

3,81E-01

7,40E-01

6,20E-01

5.5

7,17E-01

5,82E-03

2,05E-01

2,11E-01

2,04E-03

6,50E-03

2,62E-03

7,17E-01

1,32E-02

1,00E-02

7,75E-03

7,96E-03

8,03E-03

1,11E-02

4,24E-01

4,49E-01

2,62E-03

6,50E-03

2,38E-01

3,38E-01

1,00E-02

6,74E-04

1,24E-02

1,24E-01

1,22E-03

6,74E-02

5,79E-01

3,91E-01

3,85E-02

2,05E-01

1,24E-02

3,20E-01

1,12E-01

7,87E-04

7,96E-03

2,05E-01

1,00E-02

1,12E-01

6,74E-02

5,59E-01

1,59E-01

5,38E-01

5,59E-01

3,22E-02

2,11E-01

4,49E-01

4.5

8,04E-01

1,89E-02

3,52E-01

5,60E-02

1,28E-02

9,22E-03

5,98E-03

1,46E-01

1,54E-02

7,71E-03

9,22E-03

1,55E-02

5,20E-03

6,53E-02

4,36E-01

9,93E-01

4,32E-03

5,87E-02

5,56E-01

1,53E-01

3,36E-01

7,71E-03

5,98E-03

6,08E-02

2,18E-02

3,35E-02

6,07E-01

5,30E-02

5,60E-02

5,73E-02

4,95E-02

5,58E-02

1,47E-01

3,70E-03

8,08E-02

4,31E-01

5,56E-01

5,60E-02

1,61E-01

4,54E-01

4,91E-01

9,03E-01

5,20E-02

1,90E-02

2,18E-01

9,93E-01

Time (month) Time

3.5

3,50E-01

8,61E-01

4,04E-01

3,11E-01

9,57E-01

2,60E-01

3,27E-01

3,11E-01

3,27E-01

3,50E-01

3,11E-01

8,39E-01

2,60E-01

4,43E-01

9,23E-01

2,60E-01

3,50E-01

4,43E-01

3,27E-01

4,67E-01

8,26E-01

3,50E-01

3,11E-01

4,04E-01

3,11E-01

4,70E-01

9,23E-01

4,43E-01

3,11E-01

5,36E-01

9,23E-01

3,50E-01

4,56E-01

9,95E-01

3,11E-01

3,27E-01

4,04E-01

2.5

For Proteobacteria,been For has class taken into consideration

8,03E-01

5,16E-01

6,85E-01

4,72E-01

5,78E-01

9,25E-01

4,72E-01

8,75E-01

4,72E-01

6,05E-01

4,72E-01

8,03E-01

9,51E-01

7,66E-01

9,65E-01

8,03E-01

4,72E-01

5,78E-01

5,16E-01

8,03E-01

6,05E-01

4,72E-01

6,65E-01

8,03E-01

6,05E-01

4,72E-01

6,05E-01

7,33E-01

6,20E-01

4,72E-01

6,05E-01

6,37E-01

7,33E-01

4,72E-01

8,03E-01

4,72E-01

6,23E-01

5,80E-01

1.5

4,40E-01

7,30E-01

9,88E-01

4,40E-01

6,26E-01

4,40E-01

6,03E-01

7,60E-01

6,95E-01

8,23E-01

4,40E-01

6,17E-01

9,88E-01

4,40E-01

6,17E-01

7,28E-01

9,88E-01

4,40E-01

8,28E-01

6,17E-01

6,26E-01

9,52E-01

4,40E-01

6,03E-01

8,28E-01

5,49E-01

4,40E-01

Férin vs. Metal phylum occurrence comparison - FDR-correctedp-value Férin(t.test)occurrencecomparison Metalvs. phylum

0.5

0±0

0.02±0

4±0.19

0.01±0

0.04±0

0.01±0

0.02±0

0.01±0

0.02±0

0.1±0.01

0.07±0.01

0.03±0.01

0.24±0.02

0.32±0.01

0.17±0.01

0.04±0.01

0.02±0.01

0.06±0.01

0.17±0.01

0.31±0.06

7.17±0.73

6.33±0.29

4.41±0.22

0.44±0.02

0.29±0.03

0.17±0.01

0.07±0.01

8.37±0.65

5.88±0.12

8.78±0.24

0.29±0.02

0.21±0.01

3.96±0.18

3.08±0.09

4.03±0.13

12.72±0.35

10.28±0.53

17.71±0.89

Metal

6.5

0±0

0.01±0

0.03±0

0.01±0

0.02±0

0.01±0

4.6±0.48

0.6±0.06

0.15±0.03

0.02±0.01

0.11±0.02

0.35±0.02

0.14±0.01

0.05±0.01

0.03±0.01

0.02±0.01

0.05±0.01

0.11±0.02

0.24±0.02

0.06±0.01

9.19±1.04

5.86±0.35

4.07±0.23

0.51±0.03

0.12±0.02

0.15±0.01

9.34±0.94

6.78±0.17

7.84±0.18

0.35±0.03

0.24±0.01

3.41±0.12

2.95±0.08

4.44±0.16

12.73±0.49

10.35±0.69

14.98±0.94

Metal

5.5

0±0

0.01±0

0.08±0

0.01±0

0.03±0

5.6±0.06

0.49±0.02

3.64±0.21

0.06±0.01

0.25±0.02

1.58±0.14

0.33±0.01

0.12±0.02

0.17±0.02

0.16±0.02

0.06±0.01

0.02±0.01

0.19±0.02

0.54±0.02

0.04±0.01

12.37±0.5

5.38±0.29

0.02±0.01

4.35±0.22

0.52±0.03

0.07±0.01

0.14±0.02

14.5±1.24

7.81±0.44

6.41±0.23

0.67±0.06

0.27±0.02

3.33±0.34

1.95±0.07

3.74±0.17

11.34±1.46

13.73±0.92

Metal

4.5

0±0

0.02±0

0.01±0

0.04±0

0.03±0

0.01±0

0.03±0

0.2±0.02

2.4±0.16

0.35±0.04

3.86±0.27

0.07±0.01

0.25±0.03

0.78±0.05

0.32±0.01

0.17±0.02

0.12±0.01

0.11±0.01

0.06±0.01

0.11±0.01

0.65±0.02

6.37±0.66

4.75±0.23

6.33±0.78

0.56±0.06

3.62±0.19

0.06±0.01

0.17±0.02

17.4±0.87

6.35±0.18

7.05±0.16

0.67±0.06

0.24±0.01

3.73±0.12

4.61±0.23

12.65±0.29

15.86±1.63

Metal

3.5

0±0

0.01±0

0.02±0

0.01±0

0.7±0.18

2.14±0.2

7.95±0.9

0.1±0.02

5.1±0.39

1.27±0.1

0.13±0.02

0.23±0.03

0.02±0.01

0.78±0.15

0.49±0.06

0.08±0.01

0.06±0.02

0.18±0.04

0.04±0.01

0.35±0.03

0.24±0.03

0.88±0.18

0.11±0.02

6.14±0.39

0.03±0.01

3.31±0.33

3.38±0.51

0.66±0.09

1.39±0.06

0.02±0.01

4.81±0.26

0.03±0.01

1.11±0.13

0.16±0.02

2.36±0.37

3.01±0.36

17.85±0.75

34.83±2.36

Metal

2.5

0±0

0.01±0

0.03±0

0.02±0

0.01±0

0.03±0

0.4±0.06

3.68±0.6

0.2±0.02

0.1±0.02

25.5±2.2

6.5±0.64

0.11±0.02

0.02±0.01

0.85±0.26

0.64±0.09

0.16±0.02

0.04±0.01

0.09±0.01

0.46±0.07

0.23±0.03

1.42±0.27

0.07±0.01

8.55±0.96

0.24±0.12

0.01±0.01

4.96±0.31

7.65±0.85

0.79±0.08

1.78±0.25

5.57±0.25

0.98±0.06

0.12±0.01

2.03±0.16

0.82±0.06

2.94±0.08

12.84±1.33

10.03±2.87

Metal

1.5

0±0

0.02±0

0.01±0

0.02±0

3.39±0.4

0.6±0.08

0.2±0.01

1.8±0.16

3.79±0.5

2.05±0.2

6.44±0.3

0.53±0.11

0.11±0.02

0.15±0.02

0.28±0.04

0.69±0.11

0.15±0.03

0.05±0.01

0.16±0.05

0.04±0.02

0.12±0.02

0.24±0.11

4.92±0.83

0.02±0.01

5.06±1.02

0.28±0.03

0.04±0.01

0.16±0.02

5.49±0.41

0.03±0.01

0.52±0.08

0.19±0.02

1.92±0.11

1.53±0.14

2.91±0.23

13.28±2.57

12.68±3.71

30.12±3.06

Metal

0.5

0±0

0.01±0

0.4±0.03

0.2±0.02

3.17±0.2

0.2±0.02

0.07±0.02

2.36±0.22

0.03±0.01

0.09±0.03

0.05±0.01

0.41±0.03

0.01±0.01

0.08±0.02

1.78±0.16

0.05±0.01

0.02±0.01

21.9±1.15

0.06±0.02

1.48±0.14

11.5±0.65

0.04±0.01

3.04±0.17

0.05±0.01

2.72±0.21

0.04±0.01

0.02±0.01

8.98±0.99

6.84±0.41

0.08±0.02

2.09±0.22

2.84±0.14

3.51±0.31

12.98±0.66

12.85±1.15

Control

6.5

0±0

0.01±0

0.03±0

0.02±0

0.01±0

0.03±0

0.01±0

2.8±0.17

0.1±0.02

4.27±0.2

2.03±0.1

0.05±0.01

0.03±0.01

0.11±0.01

0.09±0.01

0.49±0.04

0.38±0.03

0.02±0.01

1.15±0.12

20.15±1.3

0.03±0.01

2.09±0.15

0.06±0.01

9.89±1.03

0.32±0.04

7.01±0.65

0.05±0.01

3.33±0.23

6.81±0.21

0.34±0.06

0.22±0.02

2.55±0.19

4.17±0.23

11.14±0.69

10.12±0.37

10.04±0.51

Control

5.5

0±0

0.02±0

0.01±0

0.03±0

0.02±0

0.01±0

0.03±0

0.01±0

3.2±0.18

0.59±0.1

3.9±0.13

6.6±0.11

1.32±0.1

0.15±0.03

2.32±0.06

0.03±0.01

0.05±0.01

0.12±0.01

0.11±0.01

0.27±0.02

0.24±0.01

0.04±0.01

0.14±0.01

1.37±0.03

20.6±0.91

0.03±0.01

7.14±0.16

6.33±0.28

0.34±0.02

3.94±0.15

0.05±0.01

8.27±0.34

0.51±0.05

0.28±0.02

1.55±0.08

2.83±0.15

For Proteobacteria,been For has class taken into consideration

13.97±0.53

13.49±0.47

Control

4.5

0±0

0.02±0

0.01±0

0.02±0

0.03±0

0.01±0

0.03±0

0.01±0

Férin and Metal phylum occurrence microcosm comparison (%) comparison occurrence microcosm Metal Férinphylum and

0.3±0.01

8.6±0.23

1.4±0.08

0.12±0.01

2.49±0.06

0.02±0.01

0.05±0.01

0.09±0.01

0.34±0.02

0.02±0.01

0.03±0.01

0.17±0.01

1.27±0.09

2.48±0.14

0.14±0.04

5.72±0.29

7.34±0.29

0.27±0.03

4.41±0.16

3.96±0.12

6.49±0.12

0.51±0.06

0.33±0.02

1.64±0.05

3.68±0.14

13.09±0.33

17.13±0.64

17.63±0.77

Control

3.5

0±0

0.01±0

0.9±0.17

0.1±0.02

0.42±0.1

4.64±0.9

0.35±0.11

2.47±0.65

0.07±0.03

0.07±0.02

0.03±0.01

0.13±0.02

0.18±0.03

0.02±0.01

0.06±0.03

0.02±0.01

0.18±0.06

0.13±0.03

1.35±0.34

1.72±0.23

0.01±0.01

0.02±0.01

6.04±0.53

0.02±0.01

3.08±0.73

2.55±0.16

6.06±2.45

0.01±0.01

5.56±0.85

0.02±0.01

0.86±0.17

0.19±0.02

1.33±0.19

0.87±0.14

2.29±0.25

12.84±1.79

32.91±6.31

12.44±1.93

Control

2.5

0±0

0.03±0

0.01±0

2.6±0.26

0.2±0.03

1.4±0.19

0.94±0.2

0.25±0.07

3.35±0.73

0.06±0.02

0.12±0.06

0.17±0.02

1.18±0.13

0.21±0.03

0.03±0.01

0.05±0.01

0.04±0.02

0.32±0.11

0.15±0.01

1.51±0.34

0.13±0.03

1.77±0.43

23.6±4.95

0.08±0.04

0.05±0.03

7.35±0.52

0.02±0.01

3.98±0.88

0.26±0.04

0.01±0.01

5.14±2.25

5.56±1.05

5.74±0.49

0.03±0.01

0.53±0.08

2.97±0.45

17.13±2.32

12.99±0.93

Control

1.5

0±0

0.01±0

0.56±0.1

2.53±0.1

0.2±0.03

3.86±0.4

0.26±0.1

0.46±0.13

3.72±0.34

0.12±0.02

0.12±0.03

0.02±0.01

0.16±0.02

1.37±0.16

0.27±0.04

0.04±0.01

0.09±0.01

0.05±0.01

0.15±0.02

1.52±0.26

0.06±0.04

8.81±0.44

0.01±0.01

4.49±0.54

0.45±0.11

0.01±0.01

0.02±0.01

7.31±0.59

7.89±0.63

0.05±0.01

0.77±0.26

0.23±0.03

2.11±0.25

1.58±0.14

3.63±0.21

19.56±2.04

12.76±2.35

14.69±1.31

Control

0.5

0±0

7±0.31

0.03±0

0.02±0

0.01±0

0.02±0

1±0.09

2.61±0.1

0.29±0.08

0.09±0.02

0.12±0.04

0.02±0.01

0.17±0.02

1.61±0.23

0.31±0.06

0.12±0.04

0.11±0.01

0.35±0.04

0.21±0.03

0.11±0.02

0.02±0.01

0.75±0.09

6.69±1.18

0.07±0.02

5.96±0.33

0.03±0.01

0.54±0.12

0.01±0.01

2.27±0.19

0.01±0.01

0.14±0.03

0.03±0.01

9.97±0.67

7.06±0.12

0.04±0.01

0.81±0.17

0.14±0.01

1.63±0.02

3.03±0.13

24.04±1.71

10.21±0.92

12.31±0.28

In situ In

Zea

Woesearchaeota

Verrucomicrobia

Unclassified

Thaumarchaeota

SR1

Spirochaetes

Proteobacteria

Planctomycetes

Parcubacteria

Pacearchaeota

Oligoflexia

Nitrospirae

Microgenomates

Lentisphaerae

Latescibacteria

Ignavibacteriae

Hydrogenedentes

Gemmatimonadetes

Gammaproteobacteria

Fusobacteria

Firmicutes

Euryarchaeota

Epsilonproteobacteria

Elusimicrobia

Deltaproteobacteria

Deinococcus-Thermus

Deferribacteres

Cyanobacteria/Chloroplast

Crenarchaeota

Cloacimonetes

Chloroflexi

Chlamydiae

Saccharibacteria

candidate_division_ZB3

candidate_division_WPS-1

BRC1

Betaproteobacteria

Bacteroidetes

Bacteria

Armatimonadetes

Archaea

Aminicenantes

Alphaproteobacteria

Actinobacteria

Acidobacteria Condition Time (month) Time

126

Table S5: Primers used for quantitative PCR

Target gene name Forward Reverse Reference

pbrA 5’-ACCGAAGAGGCGCTGAT-3’ 5’-GGTCGGGCAATCCATCT-3’ This study

czcA 5’-TCGACGGBGCCGTGGTSMTBGTCGAGAA-3’ 5’-GTVAWSGCCAKCGGVBGGAACA-3’ Roosa et al., 2014

IncP_OriT 5-’CAGCCTCGCAGAGCAGGAT-3’ 5’-CAGCCGGGCAGGATAGGTGAAGT-3’ Götz et al., 1996

127

Chapter 3: Selection and propagation of IncP conjugative plasmids following long-term anthropogenic metal pollution in river sediments

Journal of Hazardous Materials, 382 (2020) 121-173

Journal of Hazardous Materials 382 (2020) 121173

Contents lists available at ScienceDirect

Journal of Hazardous Materials

journal homepage: www.elsevier.com/locate/jhazmat

Selection and propagation of IncP conjugative plasmids following long-term T anthropogenic metal pollution in river sediments

Valentine Cyriaquea,b, *,1, Samuel Jacquiodb,c,1, Leise Riberd, Waleed Abu Al-soudb,e, a b,2 a,2 David C. Gillan , Søren J. Sørensen , Ruddy Wattiez a Proteomics and Microbiology Laboratory, Research Institute for Biosciences, UMONS, 20 place du parc, Mons, Belgium b Section of Microbiology, Department of Biology, University of Copenhagen, Universitetsparken 15, 2100 Copenhagen Ø, 1, Bygning, 1-1-215, Denmark c Agroécologie, UMR 1347, INRA Centre Dijon, Dijon, France d Section of Functional Genomics, Department of Biology, University of Copenhagen, Ole Maaløesvej 5, 2200 Copenhagen N, Denmark e Department of Clinical Laboratory Sciences, Faculty of Applied Medical Sciences, Jouf University, Qurayyat, Saudi Arabia

GRAPHICAL ABSTRACT

ARTICLE INFO Abstract

Editor: Deyi Hou For a century, the MetalEurop foundry released metals into the river “La Deûle”. Previous work revealed higher microbial Keywords: diversity in metal impacted sediments, and horizontal gene transfer mediated by conjugative plasmids was suggested to drive Horizontal gene transfer the community adaptation to metals. We used an integrative state-of-the-art molecular approach coupling quantitative PCR, Conjugative plasmid conjugation assays, flow cytometry, fluorescence activated cell sorting and 16S rRNA gene amplicon sequencing to investigate Metal pollution the presence of conjugative plasmids and their propagation patterns in sediment microbiomes. We highlighted the existence of River sediment a native broad-host range IncP conjugative plasmid population in polluted sediments, confirming their ecological importance FACS for microbial adaptation. However, despite incompatibilities and decreased transfer frequencies with our own alien IncP Sequencing plasmid, we evidenced that a wide diversity of bacterial members was still prone to uptake the plasmid, indicating that sediment microbial communities are still inclined to receive conjugative plasmids from the same group. We observed that metal pollution favoured exogenous plasmid transfer to specific metal-selected bacteria, which are

Abbreviations: FACS, fluorescence activated cell sorting; FRG, Functional Response Group; GFP, green fluorescent protein; HGT, horizontal gene transfer; MGE, mobile genetic element; SMCs, sediment microbial communities; OTU, operational taxonomic unit; qPCR, quantitative polymerase chain reaction; WWTP, waste-water treatment plant

*Corresponding author at: Proteomics and Microbiology Laboratory, Research Institute for Biosciences, UMONS, 20 place du parc, Mons, Belgium. E-mail address: [email protected] (V. Cyriaque). 1. These authors have contributed equally as shared-first authors. 2. These authors have contributed equally as shared-last authors.

https://doi.org/10.1016/j.jhazmat.2019.121173 Received 23 May 2019; Received in revised form 14 August 2019; Accepted 5 September 2019 Available online 06 September 2019 0304-3894/ © 2019 Elsevier B.V. All rights reserved.

128

V. Cyriaque, et al. Journal of Hazardous Materials 382 (2020) 121173 likely coming from upstream sources (e.g. wastewater treatment plant, farms…). Altogether, our results suggest that MetalEurop sediments are hotspots for gene transfer via plasmids, acting as an “environmental reservoir” for microbes and mobile elements released by human activities.

1. Introduction upcoming bacteria” (FRG3, low activity) both well-represented at Me- talEurop; the “super active faecal-related bacteria” (FRG4, very active at MetalEurop), with probable origin from upstream inlet sources (e.g. Sediment microbial communities (SMCs) are crucial players in the functioning Wastewater Treatment Plants [WWTPs] and/or agricultural origins); the of rivers (Gibbons et al., 2014). These longitudinal eco-systems are “active” (FRG5) and “passive metal sensitive bacteria” (FRG6) being impacted characterized by unilateral water-flows, directly affecting SMC assembly via by metals. This community structure was accompanied with an unexpectedly ceaseless upstream material transfers, inducing microbial community higher diversity at MetalEurop, potentially due to horizontal inheritance of coalescence processes along the flow (De Oliveira and Margis, 2015) and with metal-tolerance/resistance genes via MGE dispersion and acquisition, the hyporheic zone (Mansour et al., 2018). Beside this intrinsic dynamic aspect, especially broad-host range conjugative plasmids (Jacquiod et al., 2018a). local physico-chemical properties impact SMC structure and activity (Gibbons Moreover, the poor taxonomic relationship between metal-selected bacteria et al., 2014; Jacquiod et al., 2018a; Besemer, 2015). These peculiar properties implies an adaptive response involving broad-host range MGEs bypassing turn sediments into environmental sinks for specific compounds such as key phylogenetic barriers, like conjugative plasmids (Jacquiod et al., 2018a). Thus, elemental nutrients (e.g. N, S, P (Yuan et al., 2019; Reuther, 2009)), but also MGEs seem to be keystone features enabling the resistance and resilience of contaminants such as pesticides (Vryzas, 2018) and metals (Superville et al., these metal-stressed communities. 2014). Due to their persisting nature, metals are non-degradable and tend to accumulate in sediments where they will affect biotic entities, especially MGEs, especially plasmids (Luo et al., 2016), are genetic platforms where microorganisms that are sitting at the basis of food chains, performing crucial innovation occurs constantly through genetic re-arrangements and Horizontal ecosystem functions (e.g. nutrient cycling and bioremediation). Thus, metal Gene Transfer (HGT) (Pinilla-Redondo et al., 2018; Norman et al., 2009; pollution is regarded in ecology as an ever-lasting “press-type” disturbance, Zolgharnein et al., 2007). Conjugative plasmids are vessels for genes that may unlike “pulse-type” ones that fade away with time (Grant et al., 2017). provide a fitness-boost to their hosts, enabling fast adaptation to fluctuating Depending on environmental conditions, metals will become bioavailable and environmental conditions and/ or access to new niches (Norman et al., 2009). toxic for microorganisms, making them prone to develop resistance They are classified in incompatibility groups (i.e. plasmids from the same group mechanisms (Epelde et al., 2015) some of which being identical to the ones cannot be steadily inherited together, thus preventing simultaneous co- involved with antibiotic resistance (Pal et al., 2017). Therefore, studying the occurrence in the cell). Plasmid mobilisation in strains isolated from long-term impact of metals on microorganisms in sediments stand as an important metal-contaminated sediments was shown (Zolgharnein et al., 2007), stressing challenge in order to better understand how this ecosystem functions, and how their importance in acquisition of metal resistance genes. it may be linked to the environmental spread and maintain of harmful genes such as antibiotic resistance ones (Jacquiod et al., 2018a). Metals end up in To get a clearer picture of the environmental dispersion of conjugative sediments either from natural (Bradl, 2005) or anthropogenic sources, plasmids, one should consider “permissiveness”, the ability of a bacterial including agricultural pollutants (Alvarenga et al., 2015; Sherameti, 2015) community to receive a plasmid in term of transfer events and transconjugant leaching to rivers (Antonious et al., 2008), wastewater (Jacquiod et al., 2017), diversity (Jacquiod et al., 2017; Klümper et al., 2017). Transfer potential is smelters (Jacquiod et al., 2018a), or mining activities (Costa et al., 2015; Reis defined by the “permissive fraction”, aka bacteria able to acquire plasmids et al., 2016). Indeed, like for terrestrial ecosystems (Nayar et al., 2004; Sun et (Jacquiod et al., 2017; Sørensen et al., 2005). Still, dispersion pathways and al., 2013; Kwon et al., 2015; Berg et al., 2012; Nunes et al., 2016; Jacquiod et links with host ecology remain unclear, especially in the context of metal al., 2018b), metals are important structuring factors for the activity and pollution. Indeed, contrasting results were reported from short-term studies composition of SMCs (Jacquiod et al., 2018a; Gillan et al., 2005). assessing environmental HGT potential depending on plasmid type and donor strain (Klümper et al., 2015), as well as different stressors, like antibiotics Sediments from the Deûle river (Noyelles-Godault, France) near the abandoned (Slager et al., 2014), heat-shock (Schafer et al., 1990), SDS (Arango Pinedo MetalEurop smelter site were exposed to metal releases over 110 years, and Smets, 2005), ethanol (Seier-Petersen et al., 2014), or metals (Klümper et reaching concentrations up to 30-folds higher than ambient levels found al., 2017). This last study showed that metals decrease plasmid-uptake without upstream (Férin, France) (Gillan et al., 2015). This well-known site was modifying the transconjugant pool richness. characterized through several microbial ecology and ecotoxicology studies, showing the impact of metals on microbial and invertebrate communities Here, we used state-of-the-art molecular tools coupling conjugation assays and (Roosa et al., 2016). However, Gillan and colleagues showed that MetalEurop quantitative PCR (qPCR) to shed light on conjugative plasmids at Férin and SMCs were taxonomically and functionally similar to that of the control site MetalEurop sites. We hypothesized that conjugative plasmids were selected located upstream (Férin). Nevertheless, microbial genes involved in virulence downstream due to the metal pollution, affecting the capacity of the resident and defence mechanisms, like synthesis of protecting barriers (exopolymeric SMCs to uptake a novel alien plasmid. We quantified IncP, IncF and IncI substances), efflux systems (czcA) and Mobile Genetic Elements (MGEs), were conjugative plasmids in both sites and defined the permissiveness capacity of enriched in this metal-exposed community (Gillan et al., 2015; Roosa et al., SMCs. This study brings a considerable contribution to our understanding of 2014a). conjugative plasmid ecology in polluted sites, showing the importance of IncP plasmid on long-term resilience of metal impacted SMCs, identifying river Furthermore, using combined DNA and RNA approaches, we previously sediments as environmental transit hubs for MGEs. managed to highlight the impact of metals on the activity pat-terns of these

SMCs (Jacquiod et al., 2018a). We evidenced the existence of six distinct strategies called Functional Response Groups (FRGs, aka groups of microorganisms responding similarly to environmental cues (Jacquiod et al., 2. Experimental 2018a, 2017; Nunes et al., 2016), here metals), namely: the “dormant seed bank” (FRG1, with low activity at both sites); the “active” (FRG2, high activity) 2.1. Sediment sampling and DNA extraction and “passive Sediments were sampled in the Sensée Canal (Férin, France, 50°18′39.0″N 3°05′05.4″E) and the Deûle river (Noyelles-Godault,

2 129

V. Cyriaque, et al. Journal of Hazardous Materials 382 (2020) 121173

France, 50°25′44.7″N 3°01′20.4″E) in May 2016 (Fig. S1). The Deûle river mL). Following incubation (30 °C, 24 h), cells were harvested by vortexing flows alongside MetalEurop, an abandoned smelter that operated between filters (3 min., 2 mL, 0.9% NaCl). 1893–2003, leaching metals in the river, mainly cadmium, copper, lead and Cell detection, fluorescence counting and sorting of mCherry and GFP-emitting zinc (Table S1) (Gillan et al., 2015; Roosa et al., 2014b). Both sediments donor (red)/ transconjugant (green) cells was done using fluorescence activated display similar abiotic characteristics in the cell sorting (FACS Aria III, Becton Dickinson Biosciences, San Jose, CA, subsurface: Eh, ± 200 mV, pH, ± 8.5; dissolved organic carbon, USA). A triple-gating procedure was applied for sorting red and green −1 3− −1 −1 6–7.5 mg. L ; PO4 , 0–1 mg.L ;NH4, 0–10 mg.L +; fluorescent cells. Briefly, a pre-sorting run allowed to select (i) particles within −1 NO3, ± 0.1 mg.L (Gillan et al., 2015). the bacterial size range, followed by two consecutive enrichment runs selecting Core sampling was operated using a Plexiglas core drill (Φ =7 cm) fixed to a (ii) green fluorescent cells (488 nm laser, bandpass filter 530/30 nm) and (iii) stainless-steel bar. The first centimetre of each core was used for experiments. red fluorescent cells (561 nm laser, bandpass filter 610/20 nm) (Fig. S2). The Sampling device was washed with ethanol and air-dried before each usage. BD FACS DIVA software (v6.1.3) was used as described previously (Klümper For qPCR experiments, four sediment cores were collected from both stations et al., 2017, 2015), resulting in sorting of ∼50,000 transconjugant cells per and two samples of 2 g from core upper parts were stored at 4 °C during sample. The transfer frequencies were calculated as the ratio between transport and at −20 °C until processing. DNA extraction was performed as transconjugant cells “T” divided by donor cells “D” (Sørensen et al., 2005) described in Jacquiod et al. (2018) using 8 x 2 g of sediments per station and normalised by the total number of counted cells (B). Sorted transconjugant washed to remove PCR inhibitors as described (Fortin et al., 2004). From 2 mL (green cells) and total recipient (non-fluorescent cells) from controls were lysed washed sediments, 500 μL were used for total DNA extraction following (Lyse and Go PCR, Thermo Scientific) prior to DNA-based amplification of the manufacturer’s instructions (FastDNA® SPIN Kit for Soil DNA extraction, MP 16S rRNA gene and sequencing as described (Klümper et al., 2015). Biomedicals, Santa Ana, CA, USA).

For in vitro conjugation assays, four additional cores were collected at each 2.4. High throughput 16S rRNA gene amplicon sequencing station and two samples of 16 g from core upper parts were stored at 4 °C during transport, representing a total of 16 samples (four cores x two samples x two Lysed transconjugant and recipient cells templates served to amplify the stations, Table S2). 5 g of overnight vortexed sediments were used for cell variable V3-V4 region (approx. 460 bp) of the prokaryote 16S rRNA gene, suspensions (16 °C, settled for 10 min). including domains of archaea and bacteria using primers 341 F

2.2. Quantification of IncI, IncF and IncP plasmids by quantitative PCR (5′−CCTAYGGGRBGCASCAG-3′) and 806R (5′-GGACTACNNGGGTAT CTAAT-3′) as described (Jacquiod et al., 2017). Tagging and addition of (qPCR) sequencing adapters to amplicon were performed in a second amplification step

Three conjugative plasmid groups (IncI, IncF and IncP) were quantified by using fusion primers carrying adaptor barcode tags and spacers as described qPCR (Table S3 for primers). The reaction mixture consisted of 4 μL of (Jacquiod et al., 2018a). Purification and size-selection (removal of products < LightCycler TaqMan Master mix (Roche Life Science), 1 μL of each primer 200 bp) of the ∼620 bp amplicon were performed using Agencourt AMPure pair (10 μM), 1 μL of FAM/BHQ1-labeled probe, 12 μL of ddH20 and 2 μL of XP beads (Beckman Coulter, Brea, CA, USA) according to manufacturer’s a diluted DNA template. A LightCycler® 96 System (Roche Life Science) was instructions. Samples were pooled and adjusted to equimolar concentrations used. Amplification consisted of an initial denaturation step (95 °C, 10 min) and concentrated using the DNA Clean and Concentrator™-5 kit (Zymo followed by 40 cycles of denaturation step (95 °C, 15 s) and Research, Irvine, CA, USA). 2 × 250 bp paired-end high-throughput annealing/extension step (60 °C, 1 min). Copy numbers were inferred from sequencing on an Illumina® MiSeq® platform (Illumina, San Diego, CA, standard curves established with known quantities of each plasmid from a USA) was done according to manufacturer’s instructions. Unassembled raw private collection (Section of Microbiology, University of Copenhagen). DNA amplicon data were deposited at the Sequence Read Archive public repository concentrations were measured by Qubit™ fluorometer following (SRA, https://www.ncbi.nlm.nih.gov/sra) under the accession number manufacturer’s instructions and used to normalize copy numbers per PRJNA394660 (https://www.ncbi.nlm.nih.gov/Traces/study/?acc= microgram of DNA per gram of sediment (Table S4). SRP112522).

2.3. Solid surface filter conjugation assay 2.5. Annotation and generation of the contingency table

The capacity of SMCs to receive an exogenous conjugative plasmid (i.e. the Sequences were analysed based on best practices guidelines using qiime_pipe “permissiveness”) was tested using the in vitro conjugation assay procedure (https://github.com/maasha/qiime_pipe) as previously described (Nunes et al., previously described (Jacquiod et al., 2017; Klümper et al., 2015, 2017). 2016; Caporaso et al., 2010a; Schöler et al., 2017). Sequence demultiplexing Briefly, bacterial cell suspensions from Férin and MetalEurop sediments (n = 8 was done using the MiSeq Controller Software and diversity spacers were for each) were challenged with the donor strain Escherichia coli trimmed using biopieces (www. biopieces.org). Sequence mate-pairing and q R MG1655::lacI -Plpp-mCherry-Km , carrying an exogenous broad-host range filtering were processed using usearch v7.0.1090 (Edgar, 2010). OTU IncP-1ε plasmid, carrying a green fluorescent protein (GFP) encoding gene (Operational Taxonomic Unit) clustering, de-replication and singleton removal downstream a LacIq repressible promoter (pKJK5::Plac::gfpmut3) (Jacquiod et were performed using uparse (Edgar, 2013). Paired-end mating and trimming al., 2017; Klümper et al., 2017, 2015). Donor strain cells constitutively express were ap-plied as previously described (Jacquiod et al., 2018a). Chimera mCherry, dis-playing a red fluorescence, while sediment bacteria (i.e. removal was performed using usearch and the ChimeraSlayer package (Haas “recipients”) that acquired the conjugative plasmid (i.e. the “transconjugants”) et al., 2011). Operational Taxonomic Units (OTUs) were picked at 97% may express GFP, displaying green fluorescence. Donor and recipient cells sequence identity using Mothur v.1.25.0 (Schloss et al., 2009). An UniFrac were counted separately using combined SYBR Green and propidium iodide phylogenetic tree was built using RDP database with TrainSet9 assignation (PI) staining and flow cytometry. The conjugation device consisted in an equal with QIIME wrappers for PyNAST (Caporaso et al., 2010b), FastTree (Price et mix of donor and recipient cells onto cellulose-ester membrane filters placed al., 2009), and alignment filtering (Caporaso et al., 2010a). A read contingency on cycloheximide-supplemented (50 μg/ml) solid Luria-Bertani (LB) medium. table was exported at species level. In-formation regarding sequence counts is Controls were set using similar devices with either recipient or donor cells provided in Supporting Table S2. alone (5 × 107 bacterial cells/

3 130

V. Cyriaque, et al. Journal of Hazardous Materials 382 (2020) 121173

2.6. Statistical analyses Both recipient communities were dominated by Gammaproteobacteria and Firmicutes, with a higher proportion of Firmicutes OTUs in MetalEurop. Alpha-diversity was measured as previously described (Jacquiod et al., 2017; Gammaproteobacteria were clearly dominating the transconjugant pool of both Nunes et al., 2016). Diversity indices were calculated on rarefied data (n = stations at the expense of Firmicutes members. Alphaproteobacteria and 10,000) using eight biological replicates. The Rich-ness (OTUs) and the Bacteroidetes were well represented in transconjugants of both conditions and Shannon indices were calculated in the Past3 software (Hammer et al., 2001), Betaproteobateria were mainly found in the MetalEurop’s transconjugant pools and significance was determined with the Rgui software (v3.0.2) using (Fig. S6). multcomp (Hothorn et al., 2019) with ANOVA and a post-hoc Tukey’s HSD correction test (p < 0.05). Venn diagrams were built using limma (Ritchie et al., UniFrac multivariate analyses were used to assess the beta-diversity. Recipient 2015) based on summed and rarefied OTU profiles per treatment. Beta-diversity and transconjugant pools can be discriminated in both un-weighted (CAP1 = was assessed on rarefied data (10,000 counts) using constrained analysis of 55.0%) and weighted (CAP1 = 90.3%) UniFrac multivariate analyses. Férin principal coordinates (CAP) based on weighted and unweighted UniFrac and MetalEurop recipient communities can be dissociated using unweighted distance matrices with 1000 permutations (vegan (Dixon, 2003), picante (CAP2 = 17%) and weighted (CAP2 = 4.1%) UniFrac while the transconjugant (Kembel et al., 2010) and gunifrac (Swenson, 2014) packages). The obtained pool can only be discriminated using the weighted (CAP2 = 4.1%) UniFrac original unweighted UniFrac tree was pruned to keep transconjugants using the multivariate analysis. The interaction of both factors (Permissiveness and ape RGUI package (Paradis et al., 2004). The final pruned and unweighted Metals) is significant when considering the weighted UniFrac discriminating UniFrac tree were edited using the interactive Tree Of Life software (iTOL) analysis (Fig. 2, Table S5). (Letunic and Bork, 2016) and displays the log10 ratio between the means of transconjugant (“T”) and recipient (“R”) counts (T/R) for all transconjugants This interaction between the metal pollution and permissiveness led us to separately. Activity of dominant trans-conjugants was assessed on a heatmap investigate deeper on recipient/transconjugant distributions in Venn diagrams (Euclidian distance) using total DNA (tDNA) and total RNA (tRNA) profiles (Fig. S4C). To discern recipient OTUs that were influenced by metal, we previously obtained in-situ (Jacquiod et al., 2018a) from the same samples. The searched for OTUs that could no longer take the plasmid when extracted from log transformed RNA/DNA ratios of each OTU obtained from those total DNA MetalEurop, namely Salmonella enterica, Lysinibacillus fusiformis, and RNA profiles were confronted to the log10 ratio T/R of corresponding Lactobacillus oligofermentans, Haemophilus parainfluenzae OTUs and Pearson correlation index was measured for both Férin and Acinetobacter_sp_BN17, Unclass Caulobacteraceae, Unclass Ilumatobacter, MetalEurop SMCs. Bacillus subtilis, Vogesella indigofera, or Novosphingobium resinovorum (Table S6). We used a Pearson correlation to link the RNA/DNA ratio obtained in-situ 3. Results (Jacquiod et al., 2018a) with the T/R ratio of each transconjugant OTUs, which was significantly positive at the control site (Fig. S7A) but not in the metal- impacted community (Fig. S7B). Besides, to assess metal-impacts on plasmid 3.1. qPCR quantification of plasmid oriT and plasmid transfer rates transfer rates in diﬀerent taxa, a phylogenetic tree was constructed with iTol (Fig. 3). The permissiveness degrees in Gammaproteobacteria and We tested the occurrence of the narrow-host-range plasmids from Betaproteobacteria were heterogeneous, depending on the genus considered. incompatibility groups IncI and IncF since typical sequences belonging to some Alphaproteobacteria transconjugant mostly decreased their transfer rates in of these narrow-host-range conjugative plasmids were previously detected in metal contaminated sediments. The permissiveness of Actinobacteria members metagenomes of these sediments. We also focused on the broad-host range (e.g Gaiella, Illumatobacter Aciditeromonas,and Leucobacter) was low and plasmid IncP-1, as representatives of this group were previously isolated from decreased in MetalEurop. Likewise, the permissiveness of Bacteroidetes is polluted materials. All tested groups were detected in the sediments, but only decreased with metal contamination. Most Firmicutes are less permissive in the Inc-P plasmid copy number was significantly enriched in MetalEurop MetalEurop except Clostridiales (Fig. 3). compared to Férin (Fig. 1A). We then tested the plasmid transfer rate (T/(D*B)) of our exogenous pKJK5 (IncP-1) conjugative plasmid and we observed significantly lower transfer frequencies at MetalEurop compared to Férin (Fig.

1B). The mean transfer frequency in the river sediments was much lower than the one found in WWTPs using the exact same method (Jacquiod et al., 2017) (Fig. S3).

3.2. Comparative analysis of SMCs 16S rRNA gene profiles

The recipient and transconjugant fractions of SMCs were identified by 16S rRNA gene amplicon sequencing. The read count of each samples is displayed in table S2. Venn diagrams allowed to dispatch the observed diversity in each site for both transconjugant and recipient profiles, as well as our previous tcDNA profiles (Jacquiod et al., 2018a). Many transconjugants were not observed in the recipient pools (33% in Férin and 29% in MetalEurop; Fig. S4C). Besides, 16% (MetalEurop) and 31% (Férin) of transconjugants were not represented in their respective tcDNA profiles obtained previously (Jacquiod et al., 2018a) (Fig. S4B).

Overall, MetalEurop recipient community had higher alpha-diversity Fig. 1. Conjugative plasmid quantification (average ± SEM for Standard Error of the Mean, (Richness, ACE, Shannon and Reciprocal Simpson indices) compared to Férin. n = 5) for incompatibility groups IncF, IncI and IncP (qPCR, panel The richness in both transconjugant pools was high and not significantly 1 and transfer frequency (T/(D*B)) of our exogenous IncP-1 (average ± SEM, n = 8) diﬀerent. The Shannon and Reciprocal Simpson Indices of the transconjugant plasmid (conjugation assay, panel B). Stars indicate a significant difference (plasmid pools were proportionally decreased relatively to their respective recipient content - Kruskal-Wallis; plasmid transfer rate t-test; p-value < 0.01). fractions (Fig. S5).

4 131

V. Cyriaque, et al. Journal of Hazardous Materials 382 (2020) 121173

transconjugants” able to receive the plasmid at both sites, the “Trans-conjugants Férin” and “Transconjugants MetalEurop” (TransF and TransM) unique to each site respectively, and the “Ubiquitous fraction” not able to receive the plasmid but found in tDNA samples from both sites (Fig. 5). In the overall average

profile, when cumulating the core permissive (red) and both unique fractions (yellow and green), transconjugants represented 46% of all OTUs. Also, despite having low activity, the dormant “seed bank” (FRG1) had the highest number of core permissive and transconjugant OTUs (57%), including members from Sporichtya, Aeromonas, Shewanella, Haliscomenobacter and Sphingomonas genera which all have good plasmid permissiveness in MetalEurop. Metal-favoured FRGs (FRG2 and FRG3) get specific transconjugants in both Férin and MetalEurop, but metal-sensitive groups (FRG5, FRG6) only displayed Férin specific transconjugants (TransF). FRG4 (“Active faecal-related bacteria”) Clostridial members presented a higher transfer rate in MetalEurop.

4. Discussion

4.1. Conjugative plasmids as drivers of SMCs adaptation to long-term metal exposition

Previously, we suggested that the lack of phylogenetic relationship between bacterial members enriched at MetalEurop would imply an adaptive response through involvement of broad-host range MGEs (Jacquiod et al., 2018a). Here, we shed light on this aspect by focusing on qPCR detection of conjugative plasmids (IncF, IncI, IncP) and propagation patterns in these specific SMCs. We focused on the broad-host range plasmid IncP-1 since many representatives of this group were previously isolated from polluted material, from farming (antibiotic contaminated manure (Heuer et al., 2011; Binh et al., 2008) or pesticides treated soil (Dunon et al., 2013)) and wastewater (Akiyama et al., 2010), while also carrying catabolic and antibiotics/metal resistance genes (Popowska and Krawczyk-Balska, 2013; Jechalke et al., 2013). All tested groups were Fig. 2. Constrained Analysis of Principal Coordinates (CAP; 1000 permutations) based on present, supporting that sediments are natural environmental reservoirs for (A) unweighted (CAP1 p-value = 0.001; CAP2 p-value = 0.256) UniFrac distance or (B) conjugative plasmids. Unlike for others, IncP-1 plasmids were clearly an abundance weighted UniFrac distance (CAP1 p-value = 0.001; CAP2 p-value = 0.042) dominant as the metal pollution was likely the cause for this specific increase. discriminating recipient (unmarked cells) and transconjugant (GFP expressing cells) To further verify these quantitative molecular findings, we tested the actual fractions of Férin and MetalEurop by 16S rRNA gene amplicon sequencing carried out on transfer rate of our exogenous pKJK5 conjugative plasmid belonging to the recipients and transconjugants after filter mating assays. Variance partition associated to same incompatibility group (IncP-1), in both SMCs. Since the native IncP each model is given in Table S5. population was more abundant at MetalEurop, the establishment of an exogenous plasmid from the same incompatibility group should result in lower transfer rates. Indeed, we observed much lower transfer frequencies at In parallel, the in-situ activity patterns of transconjugants were displayed on a MetalEurop compared to Férin. Hence, the noticeable increase of native IncP heatmap (Fig. 4). Most abundant and active ones are ubiquitous OTUs, mainly plasmids in the metal-exposed SMC concurs perfectly with the reported represented by fast growing Gammaproteobacteria, despite negative metal evidence of an actively expressed metal resistome found by impacts on some species (e.g. FRG5 Klebsiella oxytoca or Pseudomonas putida metaproteogenomics on the same sites (Sun et al., 2013), likely showing the (Jacquiod et al., 2018a)). central role of broad-host conjugative plasmids in SMCs adaptation to metal. OTUs were classified into three main clusters (left-dendrogram, Fig. 4): Cluster 1 highlights active transconjugants in both sites with either high (1a) or super- high (1b) representativity. The sub-cluster 2 is formed of active species grouped Nevertheless, our data showed that transfer still occurred in MetalEurop despite into sub-clusters 2a and 2b that respectively display cosmopolitan OTUs active incompatibility, suggesting that either i) the native IncP plasmid did not spread in both sites, and metal-enriched OTUs. Sub-cluster 2b is mainly composed to all permissive cells; ii) the obtained transconjugants were freshly arrived Acinetobacter. Cluster 3 gathers rare bacteria with a higher representation of newcomers that had not yet acquired an IncP plasmid; iii) some carrying cells Alpha-proteobacteria, Firmicutes and Betaproteobacteria. Cluster 3 includes have lost the plasmid over time. The plasmid-host association is a dynamic the sub-cluster 3a made of metal-enriched OTUs and sub-cluster 3b made of process (Pinilla-Redondo et al., 2018) governed by trade-oﬀs between OTUs that were quasi-absent in our previous DNA/RNA-based study (Jacquiod advantages they may procure (Burlage et al., 1989; Monchy et al., 2007), their et al., 2018a), but still detected here thanks to FACS sensitivity. maintenance strategy (e.g. toxin/antitoxin systems (Hülter et al., 2017)), and their metabolic cost which hampers their persistence in microbial communities (Carroll and Wong, 2018). Therefore, one could hypothesize that plasmid-loss 3.3. Link the permissiveness with FRGs of the communities may be favoured over time due to costly energy requirements for both replication/transfer (San Millan and MacLean, 2017; Gama et al., 2018), Complementing our six FRGs (Jacquiod et al., 2018a), we defined four accessory genes expression and metal detoxification. Moreover, the option of categories in transconjugant pools to establish links between activity and being a cheater sneaking on extra-cellular “public goods” provided by other permissiveness (Fig. 5). We defined the “Core permissive bacteria, especially

5 132

V. Cyriaque, et al. Journal of Hazardous Materials 382 (2020) 121173

Fig. 3. UniFrac phylogenetic tree of transconjugants obtained from 16S rRNA gene amplicon sequencing carried out on transconjugants (GFP expressing cells) after filter mating assays on Férin and MetalEurop sediments. Turquoise to purple gradient represents the estimated transfer rate of the exogenous IncP-1ε plasmid for each OTU, calculated as the log10 ratio between transconjugant cells of that OTU (T) divided by the number of recipient cells of the same OTU (R). The inner and outer heatmap circles are respectively indicating the transfer rates at Férin and MetalEurop. The profile of these transconjugant OTUs obtained from tRNA and tDNA counts from 16S rRNA gene amplicon sequencing from Férin and MetalEurop by Jacquiod et al. (2018) is displayed in Fig. 4. the likes protecting against metal (O’Brien et al., 2014), would further favour 4.2. Comparative analysis of SMCs permissiveness plasmid-loss, making cells free for the acquisition of new plasmids from the same incompatibility group. Indigenous cells available for receiving a new Discrepancies between our previous 16S rRNA profiles (Jacquiod et al., 2018a) conjugative plasmid may acquire it via the local native population, or via any and our transconjugants/recipients fractions identified here from both stations donor-cell coming from upstream with the water flow such as from WWTP were previously observed (Jacquiod et al., 2017), likely due to methodological outlets (Akiyama et al., 2010) or leaching from farming sources (Binh et al., aspects such as (i) the vortex-extraction of cells before conjugation and (ii) the 2008; Dunon et al., 2013). Indeed, it is important to note that a WWTP and over-growth of heterotrophic fast-growing recipient cells during the overnight farms are located between Férin and MetalEurop, and that our previous FRG incubation, evidenced by Gammaproteobacteria dominance (Fig. S5, (Epelde analysis pinpointed the potential metal-selection of such allochthonous bacteria et al., 2015; Burlage et al., 1989)). Still, transconjugant OTUs were absent from from upstream water flows and connected sources (FRG4: “Super active faecal- the tcDNA profiles, highlighting the higher FACS sensitivity over MiSeq related bacteria” (Jacquiod et al., 2018a)). Besides, a well-known co-selection sequencing for the detection of infrequent events, like plasmid transfers to rare process between metal and antibiotic resistance genes (ARGs) occur in the taxa. environment (Baker-Austin et al., 2006). Consequently, the plasmid-turnover in SMCs may open potential propagation routes for harmful genes encoding The high diversity found in MetalEurop recipient concurs with our previous ARGs coming from WWTPs or agricultural hotspots. The large pool of IncP observations (Jacquiod et al., 2018a). Likewise, as previously shown (Luo et plasmids coupled to the low transfer frequencies observed in MetalEurop al., 2016), metal does not influence transconjugal pool richness, as no diﬀerence clearly indicate that there are now fewer permissive bacteria left, and that was observed between Férin and MetalEurop. A decrease in evenness (Shannon plasmid expansion was likely achieved through HGT rather than by radial and Reciprocal Simpson Indices) of the transconjugant profiles indicated replication. Thus, these findings emphasise the importance of HGT processes uneven distribution due to the dominance of some transconjugants (e.g. in environmental bacterial community adaptation, which is yet rarely observed Gammaproteobacteria) as seen previously (Li et al., 2018). This was also noted to such a fine scale. on beta-diversity, revealing overall similar taxonomic profiles for Férin and MetalEurop transconjugant pools compared to recipients. As for alpha- diversity, the

6 133

V. Cyriaque, et al. Journal of Hazardous Materials 382 (2020) 121173

Fig. 4. Log transformed representation in DNA (Férin: light blue; MetalEurop dark blue) and RNA (Férin: light pink; MetalEurop dark pink) of the selected dominant transconjugants found in this study (∑replicates OTUs counts in sorted-transconjugant > 5). Activity patterns were adapted from Jacquiod et al., 2018, and plotted in a heatmap (Euclidean distance, average clustering). Data were centred and scaled by sample (per column average). Numbered nodes correspond to species graduated by occurrence, being abundant [1a], the most abundant [1b], common species [2a], common and enriched in MetalEurop [2b], rare and enriched in MetalEurop [3a] and rare and quasi-absent in tDNA [3b].

7 134

V. Cyriaque, et al. Journal of Hazardous Materials 382 (2020) 121173

Fig. 5. Distribution of transconjugants obtained in this study in the pre-defined FRGs established from RNA/DNA ratios (FRGs) (Jacquiod et al., 2018a). Data obtained from Jacquiod et al. 2018 was re-used to determine which OTUs in each FRGs were able to receive the exogenous conjugative plasmid IncP-1ε. OTU members of FRGs (FRG1, n = 36; FRG2, n = 29; FRG3, n = 39; FRG3, n = 13; FRG5, n = 11; FRG6, n = 11) and the “average profile” (distribution of transconjugant in the entire microbial community) were classified based on their appearance in the transconjugant profiles of Férin and MetalEurop obtained in this study as follow: [25] “Core Permissive” if OTUs were transconjugant in both Férin and MetalEurop, (ii) “TranconjF” displays transconjugants unique at Férin, while (iii) “TransconjM” gives transconjugants unique at MetalEurop, (iv) “Ubiquitous” are not transconjugants but are OTUs present in both Férin and MetalEurop tDNA samples, and (v) “Others” are remaining FRG OTUs that couldn’t be assigned in our transconjugant dataset but are present in-situ. “Core Permissives”, “TransconjF” and “TransconjM” represent the total amount of bacteria that acquired the exogenous conjugative plasmid in each FRG. diﬀerence between both sites was not marked, especially for tranconjugant Gammaproteobacteria was heterogeneous, depending on the genus considered pools where an eﬀect was noticed only on weighted UniFrac, denoting some and activity. Most abundant and active ones (Cluster 1b: Citrobacter and an homogeneity degree in plasmid permissiveness between sites due to plasmid unclassified Enterobacter) have decreased transfer rate in MetalEurop, while preferential host-scope (Jacquiod et al., 2017; Li et al., 2018). Indeed, as dominant strains of cluster 1a (e.g. Aeromonas hydrophila, Pseudomonas Firmicutes were favoured at the expense of Gammaproteobacteria in putida or Klebsiella oxytoca) and abundant strains of cluster 2a (e.g. Roultella MetalEurop recipient fractions, their representation receded to even levels in plancticola, Aeromonas media or Pseudomonas syringe) show an increased transconjugant pools (Fig. S5). Besides, as previously noted (Jacquiod et al., transfer rate with metals. Metal-enriched Acinetobacter (FRG2: the metal- 2017; Li et al., 2018), this plasmid easily transfers to Alphaproteobacteria, active “upcoming bacteria” (Jacquiod et al., 2018a), cluster 2b) also displays a being enriched in transconjugant pools of both stations compared to recipients. higher plasmid transfer rate in metal contaminated sediments. Rare Aeromonas By deduction, cosmopolitan bacteria that could no longer receive the plasmid species (cluster 3) displayed a decreased transfer rate in MetalEurop. The in MetalEurop most likely already carried an indigenous IncP plasmid. Indeed, Betaproteobacteria transconjugants (e.g. Delftia acidivorans) mostly belongs we were able to pinpoint eleven of such OTUs found in recipients of both sites, to the cluster 3 “rare bacteria” but there is no clear metal-related trend on their but which could only take the plasmid in Férin and not anymore in MetalEurop, permissiveness. Most Alphaproteobacteria transconjugants have decreased indicating that they may be the ones already carrying the native IncP plasmids. transfer rates in metal contaminated sediments. The permissiveness of Some of them are known residents of other habitats such as in WWTPs for Actinobacteria and Bacteroidetes members are low and are likely negatively Salmonella enterica and Lysinibacillus fusiformis (He et al., 2011; Masarikova impacted by metals. Most gram-positive Firmicutes are less permissive in et al., 2003), food environments such as Lactobacillus oligofermentans (Rouger MetalEurop but super-active Clostridiales transconjugants (FRG4) (Jacquiod et al., 2017) or medical environments such as Haemophilus parainfluenzae et al., 2018a), display an increased permissiveness with metal contamination. (Pang and Swords, 2017; Deza et al., 2016). These findings thus reinforce the They have already been identified as good permissive strains in WWTP context notion that SMCs can be considered as an important environmental hotspot for (Jacquiod et al., 2017) and under metal stress (Klümper et al., 2017). Although HGT involving bacteria from various anthropogenic sources (Jacquiod et al., rare, transfer between gram-positive and negative is possible (Courvalin, 1994) 2018a; Epelde et al., 2015). due to secretion mechanism similarities (Goessweiner-mohr et al., 2014), especially between Firmicutes and Proteobacteria (Kintses et al., 2019). The

natural competence of some Firmicutes species could also explain plasmid 4.3. Linking strain permissiveness and activity uptake (Jacquiod et al., 2017; Johnsen et al., 2009), which may be enhanced by metals like for other stressors (antibiotic treatment, DNA damage (Blokesch, Overall, a weak but significant positive correlation between in-vitro plasmid 2016)). Besides, other genera containing harmful pathogen representatives transfer and in-situ activity of each OTU indicates that microbial activity is engage more plasmid transfer in MetalEurop sediments, such as Sta- indeed a key factor for HGT. This trend vanished in the MetalEurop, likely due phylococcus, Streptococcus, Pseudomonas, Aeromonas, Shewanella, and to the metal selection for more IncP plasmids and the resulting higher activity Acinetobacter which were already reported for their permissiveness in WWTPs observed compared to Férin (Jacquiod et al., 2018a). This observation (Jacquiod et al., 2017; Li et al., 2018). Therefore, MetalEurop SMCs are likely reinforces the link between conjugative plasmid presence and activity, further acting as transfer hubs for IncP conjugative plasmids from anthropogenic confirming that the bacterial community has successfully adapted to the sources to the environment. pollution. The in-situ activity patterns (Fig. 4) revealed a gradient of activity amongst transconjugants. Activity-based clusters revealed phylogenetic trends associated with the plasmid uptake capacity displayed on the phylogenetic tree and with metal pollution. Indeed, the permissiveness degree in

135

V. Cyriaque, et al. Journal of Hazardous Materials 382 (2020) 121173

4.4. Activity patterns revisited with HGT capacities Acknowledgements

Globally, the fraction of transconjugant OTUs from the overall community is coherent with previous results (Jacquiod et al., 2017). Besides, while metal- Authors wish to express a deep gratitude to Luma George Odish and all students favoured OTUs (FRGs 3 and 4) count transconjugants among both Férin and that have attended the 2016 master course “Emerging Molecular Techniques in MetalEurop SMCs, metal-sensitive groups (FRGs 5 and 6) displayed only Férin Microbiology” provided by the Section of Microbiology at the University of specific transconjugants. Those last OTUs may already have acquired an IncP Copenhagen for their participation to the lab work and lectures associated with plasmid at MetalEurop and may be inapt to host a new exogenous IncP plasmid. this research project. The combined burden of metals and the carried plasmid would explain why the activity of those bacteria is decreased in MetalEurop sediments, high-lighting Supplementary data the settlement of IncP plasmids in the metal-impacted com-munity. Supplementary material related to this article can be found, in the online Gammaproteobacteria from the active metal sensitive bacteria (FRG5; version, at doi:https://doi.org/10.1016/j.jhazmat.2019.121173. Pseudomonas putida, Klebsiella oxytoca and an unclassidied Co- mamonadaceae) are permissive in both sites, indicating that those metal- References sensitive bacteria could still acquire plasmid-mediated metal-resistance in Akiyama, T., Asfahl, K.L., Savin, M.C., 2010. Broad-host-range plasmids in treated was-tewater eﬄuent and MetalEurop. In line with our previous findings, active faecal-related bacteria receiving streams. J. Environ. Qual. 39, 2211–2215.

Clostridial members presented a higher transfer rate in MetalEurop, suggesting Alvarenga, P., Mourinha, C., Farto, M., Santos, T., Palma, P., Sengo, J., Morais, M.C., Cunha-Queda, C., 2015. their facilitated settlement from potential upstream sources (e.g. WWTPs, Sewage sludge, compost and other representative organic wastes as agricultural soil amendments: benefits versus limiting factors. Waste Manage. 40, 44–52. farms) in metal-contaminated sediments (Jacquiod et al., 2018a). These results Antonious, G.F., Turley, E.T., Sikora, F., Snyder, J.C., 2008. Heavy metal mobility in runoff water and show the real potential of these allochthonous members from upstream absorption by eggplant fruits from sludge treated soil. J. Environ. Sci. Health - Part B Pestic. Food Contam. sediments to engage in HGT events and acquire/exchange MGEs in Agric. Wastes 43, 526–532. downstream metal-polluted sediments. Consequently, these types of bacteria Arango Pinedo, C., Smets, B.F., 2005. Conjugal TOL transfer from Pseudomonas putida to Pseudomonas may be shuttles for the environmental acquisition, maintenance and diffusion aeruginosa: effects of restriction proficiency, toxicant exposure, cell density ratios, and conjugation detection method on observed transfer efficiencies. Appl. Environ. Microbiol. 71, 51–57. of MGEs, while potentially carrying ARGs. Narrow-host range plasmids or Baker-Austin, C., Wright, M.S., Stepanauskas, R., McArthur, J.V., 2006. Co-selection of antibiotic and metal viral genes may also be a source of MRGs and ARGs and should also be resistance. Trends Microbiol. 14, 176–182. investigated as many elements related to phages were found exclusively in Berg, J., Brandt, K.K., Al-Soud, W.A., Holm, P.E., Hansen, L.H., Sørensen, S.J., Nybroe, O., 2012. Selection for MetalEurop metagenomes (Gillan et al., 2015), and as phages are known to Cu-tolerant bacterial communities with altered composition, but unaltered richness, via long-term cu exposure. play an important role in microbial communities coping with metals (Jacquiod Appl. Environ. Microbiol. 78, 7438–7446. et al., 2018b) while acting as vehicles for antibiotic resistance in wastewater Besemer, K., 2015. Biodiversity, community structure and function of biofilms in stream ecosystems. Res. Microbiol. 166, 774–781. (Gunathilaka et al., 2017). Binh, C.T.T., Heuer, H., Kaupenjohann, M., Smalla, K., 2008. Piggery manure used for soil fertilization is a reservoir for transferable antibiotic resistance plasmids. FEMS Microbiol. Ecol. 66, 25–37.

Blokesch, M., 2016. Natural competence for transformation. Curr. Biol. 26, R1126–R1130. 5. Conclusion Bradl, H.B., 2005. Sources and origins of heavy metals. In: Bradl, H.B. (Ed.), Heavy Met. These novel molecular findings are significantly reinforcing our fundamental Environ. Elsevier Ltd., pp. 1–27. knowledge on the ecology of broad-host conjugative plasmids in metal- Burlage, R.S., Hooper, S.W., Sayler, G.S., 1989. The TOL (pWWO) catabolic plasmid. polluted environments, showing, for the first time, that enhancement of IncP Appl. Environ. Microbiol. 55, 1323–1328. plasmids in long-term metal-contaminated sediments was most likely carried Caporaso, J.G., Justin, K., Jesse, S., Kyle, B., Frederic, D.B., Elizabeth, K.C., Noah, F., Antonio, G.P., Julia, K.G., Jeﬀrey, I.G., Gavin, A.H., Scott, T.K., Dan, K., Jeremy, E.K., Ruth, E.L., Catherine, A.L., Daniel, M., through HGT. These results are supporting our previous postulates Brian, D.M., Meg, P., Jens, R., Joel, R.S., Peter, J.T., William, A.W., Widmann, J., Tanya, Y., Jesse, Z., Rob, K., hypothesizing that MetalEurop SMCs could act as an environmental hub for 2010a. QIIME allows analysis of high-throughput community sequencing data. Nat. Methods 7, 335–336. plasmid exchange. Results support the dynamic aspect of environmental Caporaso, J.G., Bittinger, K., Bushman, F.D., Desantis, T.Z., Andersen, G.L., Knight, R., 2010b. PyNAST: a plasmid turnover, being open for exogenous supplies via water streams flexible tool for aligning sequences to a template alignment. Bioinformatics 26, 266–267. potentially coming from anthropized sources such as WWTPs or farms. Carroll, A.C., Wong, A., 2018. Plasmid persistence: costs, benefits, and the plasmid paradox. Can. J. Microbiol. 64, 293–304. Allochthonous bacteria coming from anthropogenic sources are depicted as highly permissive, spanning a diverse spectrum of taxa including rare species. Costa, P.S., Reis, M.P., Ávila, M.P., Leite, L.R., De Araújo, F.M.G., Salim, A.C.M., Oliveira, G., Barbosa, F., Chartone-Souza, E., Nascimento, A.M.A., 2015. Metagenome of a microbial community inhabiting a metal-rich To get a better picture of MGE dynamics, additional work going beyond the tropical stream sediment. PLoS One 10, 1–21. scope of this study should be performed to verify the correlation between metal Courvalin, P., 1994. Transfer of antibiotic resistance genes between gram-positive and gram-negative bacteriat. pollution and conjugative plasmid presence at a much broader scale. Antimicrob. Agents Chemother. 38, 1447–1451.

Furthermore, narrow-host range plasmids and viral genes should also be taken De Oliveira, L.F.V., Margis, R., 2015. The source of the river as a nursery for microbial diversity. PLoS One into consideration, as they may also be involved into lateral transfer of 10, 1–11. antibiotic and metal resistance genes. Deza, G., Martin-Ezquerra, G., Gómez, J., Villar-García, J., Supervia, A., Pujol, R.M., 2016. Isolation of haemophilus influenzae and haemophilus parainfluenzae in ure-thral exudates from men with acute urethritis: a descriptive study of 52 cases. Sex. Transm. Infect. 92, 29–31.

Dixon, P., 2003. VEGAN, a package of R functions for community ecology. J. Veg. Sci. 14, 927–930.

Funding Dunon, V., Sniegowski, K., Bers, K., Lavigne, R., Smalla, K., Springael, D., 2013. High prevalence of IncP-1 This research was funded by the ITN Marie-Curie project Train-biodiverse n◦ plasmids and IS1071 insertion sequences in on-farm biopur-ification systems and other pesticide-polluted environments. FEMS Microbiol. Ecol. 86, 415–431. REA 289949 (SJ) (European commission), by the Fund for Scientific Research Edgar, R.C., 2010. Search and clustering orders of magnitude faster than BLAST. (F.R.S-FNRS) FRFC 7050357, PF014F828/A09016F and T.0127.14 Bioinformatics 26, 2460–2461. (Belgium), by the Pole d’ Attraction Inter-universitaire (PAI) n◦ P7/25 (VC, Edgar, R.C., 2013. UPARSE: Highly accurate OTU sequences from microbial amplicon reads. Nat. Methods DG, RW) (Belgium), and by the Center for Environmental and Agricultural 10, 996–998.

Microbiology (CREAM2) funded by The Villum Foundation (LR) (Denmark). Epelde, L., Lanzén, A., Blanco, F., Urich, T., Garbisu, C., 2015. Adaptation of soil mi-crobial community All cited funding played a role in each step of the research. structure and function to chronic metal contamination at an abandoned Pb-Zn mine. FEMS Microbiol. Ecol. 91, 1–11. Fortin, N., Beaumier, D., Lee, K., Greer, C.W., 2004. Soil washing improves the recovery of total community DNA from polluted and high organic content sediments. J. Microbiol. Methods 56, 181–191.

Gama, J.A., Zilhão, R., Dionisio, F., 2018. Impact of plasmid interactions with the chro-mosome and other plasmids on the spread of antibiotic resistance. Plasmid 99, 82–88.

136

V. Cyriaque, et al. Journal of Hazardous Materials 382 (2020) 121173

Gibbons, S.M., Jones, E., Bearquiver, A., Blackwolf, F., Roundstone, W., Scott, N., Hooker, J., Madsen, R., Nayar, S., Goh, B.P.L., Chou, L.M., 2004. Environmental impact of heavy metals from dredged and resuspended Coleman, M.L., Gilbert, J.A., 2014. Human and environmental im-pacts on river sediment microbial sediments on phytoplankton and bacteria assessed in in-situ mesocosms. Ecotoxicol. Environ. Saf. 59, 349–369. communities. PLoS One 9, 1–9. Norman, A., Hansen, L.H., Sørensen, S.J., 2009. Conjugative plasmids: vessels of the communal gene pool. Gillan, D.C., Danis, B., Pernet, P., Joly, G., Dubois, P., 2005. Structure of sediment-as-sociated microbial Philos. Trans. R. Soc. B Biol. Sci. 364, 2275–2289. communities along a heavy-metal contamination gradient in the marine environment. Appl. Environ. Microbiol. 71, 679–690. Nunes, I., Jacquiod, S., Brejnrod, A., Holm, P.E., Johansen, A., Brandt, K.K., Priemé, A., Sørensen, S.J., 2016. Coping with copper: legacy eﬀect of copper on potential activity of soil bacteria following a century of exposure. Gillan, D.C., Roosa, S., Kunath, B., Billon, G., Wattiez, R., 2015. The long-term adaptation of bacterial FEMS Microbiol. Ecol. 92, 1–12. communities in metal-contaminated sediments: a metaproteogenomic study. Environ. Microbiol. 17, 1991–2005. O’Brien, S., Hodgson, D.J., Buckling, A., 2014. Social evolution of toxic metal bior-emediation in Goessweiner-mohr, N., Arends, K., Keller, W., Grohmann, E., 2014. Conjugation in gram-positive bacteria. Pseudomonas aeruginosa. Proc. R. Soc. B Biol. Sci. 281 20140858–20140858. Microbiol. Spectr. 2, 1–19. Pal, C., Asiani, K., Arya, S., Rensing, C., Stekel, D.J., Larsson, D.G.J., Hobman, J.L., 2017. Grant, P.R., Schmitt, J., Grant, B.R., Huey, R.B., Johnson, M.T.J., Knoll, A.H., 2017. Metal Resistance and Its Association With Antibiotic Resistance, 1st ed. Elsevier Ltd. Evolution caused by extreme events. Philos. Trans. R. Soc. B Biol. Sci. 372, 20160146. Pang, B., Swords, E., 2017. Haemophilus parainfluenzae strain ATCC 33392 forms bio-films in vitro and during Gunathilaka, G.U., Tahlan, V., Mafiz, A.I., Polur, M., Zhang, Y., 2017. Phages in urban wastewater have the experimental otitis media infections. Infect. Immun. 85, 1–13. potential to disseminate antibiotic resistance. Int. J. Antimicrob. Agents 50, 678–683. Paradis, E., Claude, J., Strimmer, K., 2004. APE: Analyses of phylogenetics and evolution in R language. Haas, B.J., Gevers, D., Earl, A.M., Feldgarden, M., Ward, D.V., Giannoukos, G., Ciulla, D., Tabbaa, D., Bioinformatics 20, 289–290. Highlander, S.K., Sodergren, E., Methé, B., DeSantis, T.Z., The Human Microbiome Consortium, Petrosino, J.F., Knight, R., Birren, B.W., 2011. Chimeric 16S rRNA sequence formation and detection in sanger and 454- Pinilla-Redondo, R., Cyriaque, V., Jacquiod, S., Sørensen, S.J., Riber, L., 2018. Monitoring plasmid-mediated pyrosequenced PCR am-plicons. Genome Res. 21, 494–504. horizontal gene transfer in microbiomes: recent advances and fu-ture perspectives. Plasmid 99, 56–67.

Hammer, Ø., Harper, D.A.T.A.T., Ryan, P.D., 2001. PAST: paleontological statistics soft-ware package for Popowska, M., Krawczyk-Balska, A., 2013. Broad-host-range IncP-1 plasmids and their resistance potential. education and data analysis. Palaeontol. Electron. 4 (1), 1–9. Front. Microbiol. 4. 1-8

He, M., Li, X., Liu, H., Miller, S.J., Wang, G., Rensing, C., 2011. Characterization and Price, M.N., Dehal, P.S., Arkin, A.P., 2009. Fasttree: Computing large minimum evolution trees with profiles genomic analysis of a highly chromate resistant and reducing bacterial strain Lysinibacillus fusiformis instead of a distance matrix. Mol. Biol. Evol. 26, 1641–1650. ZC1. J. Hazard. Mater. 185, 682–688. Reis, M.P., Dias, M.F., Costa, P.S., Ávila, M.P., Leite, L.R., de Araújo, F.M.G., Salim, A.C.M., Bucciarelli- Heuer, H., Schmitt, H., Smalla, K., 2011. Antibiotic resistance gene spread due to manure application on Rodriguez, M., Oliveira, G., Chartone-Souza, E., Nascimento, A.M.A., 2016. Metagenomic signatures of a agricultural fields. Curr. Opin. Microbiol. 14, 236–243. tropical mining-impacted stream reveal complex microbial and metabolic networks. Chemosphere 161, 266– 273. Hothorn, T., Bretz, F., Westfall, P., Heiberger, R.M., Schuetzenmeister, A., Scheibe, S., 2019. The “multcomp” package. In: Bretz, F., Hothorn, T., Westfall, P. (Eds.), Mult. Comp. Using R. CRC Press, Taylor & Francis Reuther, R., 2009. Lake and River Sediment Monitoring, Encyclopedia of Life Support Systems (EOLSS). Group, pp. 1–36. https://cran.r-project. org/web/packages/multcomp/multcomp.pdf. Ritchie, M.E., Phipson, B., Wu, D., Hu, Y., Law, C.W., Shi, W., Smyth, G.K., 2015. Limma powers diﬀerential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res. 43, e47. Hülter, N., Ilhan, J., Wein, T., Kadibalban, A.S., Hammerschmidt, K., Dagan, T., 2017. An evolutionary perspective on plasmid lifestyle modes. Curr. Opin. Microbiol. 38, 74–80. Roosa, S., Wattiez, R., Prygiel, E., Lesven, L., Billon, G., Gillan, D.C., 2014a. Bacterial metal resistance genes and metal bioavailability in contaminated sediments. Environ. Pollut. 189, 143–151. Jacquiod, S., Brejnrod, A., Morberg, S.M., Abu Al-Soud, W., Sørensen, S.J., Riber, L., 2017. Deciphering conjugative plasmid permissiveness in wastewater microbiomes. Mol. Ecol. 26, 3556–3571. Roosa, S., Wattiez, R., Prygiel, E., Lesven, L., Billon, G., Gillan, D.C., 2014b. Bacterial metal resistance genes and metal bioavailability in contaminated sediments. Environ. Pollut. 189, 143–151. Jacquiod, S., Cyriaque, V., Riber, L., Al-soud, W.A., Gillan, D.C., Wattiez, R., Sørensen, S.J., 2018a. Long-term industrial metal contamination unexpectedly shaped diversity and activity response of sediment microbiome. J. Roosa, S., Prygiel, E., Lesven, L., Wattiez, R., Gillan, D., Ferrari, B.J.D., Criquet, J., Billon, G., 2016. On the Hazard. Mater. 344, 299–307. bioavailability of trace metals in surface sediments: a combined geochemical and biological approach. Environ. Sci. Pollut. Res. 23, 10679–10692. Jacquiod, S., Nunes, I., Brejnrod, A., Hansen, M.A., Holm, P.E., Johansen, A., Brandt, K.K., Priemé, A., Sørensen, S.J., 2018b. Long-term soil metal exposure impaired temporal variation in microbial Rouger, A., Tresse, O., Monique, Z., 2017. Bacterial contaminants of poultry meat: sources, species, and dynamics. Microorganisms 5, 1–16. metatranscriptomes and enriched active phages. Microbiome 6, 1–14. Jechalke, S., Dealtry, S., Smalla, K., Heuer, H., 2013. Quantification of IncP-1 plasmid prevalence in environmental: samples. Appl. Environ. Microbiol. 79, 1410–1413. San Millan, A., MacLean, R.C., 2017. Fitness costs of plasmids: a limit to plasmid trans-mission. Microbiol. Spectr. 5, 1–12. Johnsen, P.J., Dubnau, D., Levin, B.R., 2009. Episodic selection and the maintenance of competence and natural transformation in Bacillus subtilis. Genetics 181, 1521–1533. Schafer, A., Kalinowski, J., Simon, R., Puhler, A., 1990. High-frequency conjugal plasmid transfer from gram- negative Escherichia coli to various gram-positive coryneform bacteria. J. Bacteriol. 172, 1663–1666. Kembel, S.W., Cowan, P.D., Helmus, M.R., Cornwell, W.K., Morlon, H., Ackerly, D.D., Blomberg, S.P., Webb, C.O., 2010. Picante: R tools for integrating phylogenies and ecology. Bioinformatics 26, 1463–1464. Schloss, P.D., Westcott, S.L., Ryabin, T., Hall, J.R., Hartmann, M., Hollister, E.B., Lesniewski, R.A., Oakley, B.B., Parks, D.H., Robinson, C.J., Sahl, J.W., Stres, B., Thallinger, G.G., Van Horn, D.J., Weber,

Kintses, B., Méhi, O., Ari, E., Számel, M., Györkei, Á., Jangir, P.K., Nagy, I., Pál, F., Fekete, G., Tengölics, R., C.F., 2009. Introducing mothur: open-source, platform-independent, community-supported software for Nyerges, Á., Likó, I., Bálint, A., Molnár, T., Bálint, B., Vásárhelyi, B.M., Bustamante, M., Papp, B., Pál, C., describing and comparing microbial communities. Appl. Environ. Microbiol. 75, 7537–7541. 2019. Phylogenetic barriers to horizontal transfer of antimicrobial peptide resistance genes in the human. Nat. Microbiol. 4, 447–458. Schöler, A., Jacquiod, S., Vestergaard, G., Schulz, S., Schloter, M., 2017. Analysis of soil microbial communities based on amplicon sequencing of marker genes. Biol. Fertil. Soils 53, 485–489.

Klümper, U., Riber, L., Dechesne, A., Sannazzarro, A., Hansen, L.H., Sørensen, S.J., Smets, B.F., 2015. Broad Seier-Petersen, M.A., Jasni, A., Aarestrup, F.M., Vigre, H., Mullany, P., Roberts, A.P., Agersø, Y., 2014. Eﬀect host range plasmids can invade an unexpectedly diverse fraction of a soil bacterial community. ISME J. 9, 934– of subinhibitory concentrations of four commonly used biocides on the conjugative transfer of Tn916 in 945. Bacillus subtilis. J. Antimicrob. Chemother. 69, 343–348.

Klümper, U., Dechesne, A., Riber, L., Brandt, K.K., Gülay, A., Sørensen, S.J., Smets, B.F., 2017. Metal stressors Sherameti, I., 2015. Heavy Metal Contamination of Soils. Springer International Publishing. consistently modulate bacterial conjugal plasmid uptake po-tential in a phylogenetically conserved manner. ISME Slager, J., Kjos, M., Attaiech, L., Veening, J.W., 2014. Antibiotic-induced replication stress triggers bacterial J. 11, 152–165. competence by increasing gene dosage near the origin. Cell 157, 395–406.

Kwon, M.J., Yang, J.S., Lee, S., Lee, G., Ham, B., Boyanov, M.I., Kemner, K.M.,O’Loughlin, E.J., 2015. Geochemical characteristics and microbial community composition in toxic metal-rich sediments contaminated Sørensen, S.J., Bailey, M., Hansen, L.H., Kroer, N., Wuertz, S., 2005. Studying plasmid horizontal transfer in- with Au-Ag mine tailings. J.Hazard. Mater. 296, 147–157. situ: a critical review. Nat. Rev. Microbiol. 3, 700–710.

Letunic, I., Bork, P., 2016. Interactive tree of life (iTOL) v3: an online tool for the display and annotation of Sun, M.Y., Daﬀorn, K.A., Johnston, E.L., Brown, M.V., 2013. Core sediment bacteria drive community response phylogenetic and other trees. Nucleic Acids Res. 44, W242–W245. to anthropogenic contamination over multiple environmental gradients. Environ. Microbiol. 15, 2517–2531.

Li, L., Dechesne, A., He, Z., Madsen, J.S., Nesme, J., Sorensen, S.J., Smets, B.F., 2018. Estimating the Superville, P.-J., Prygiel, E., Magnier, A., Lesven, L., Gao, Y., Baeyens, W., Ouddane, B., Dumoulin, D., Billon, transfer range of plasmids encoding antimicrobial resistance in a wastewater treatment plant microbial G., 2014. Daily variations of Zn and Pb concentrations in the Deûle River in relation to the resuspension of heavily community. Environ. Sci. Technol. Lett ac-s.estlett.8b00105. polluted sediments. Sci. Total Environ. 470–471, 600–607.

Luo, W., Xu, Z., Riber, L., Hansen, L.H., Sørensen, S.J., 2016. Diverse gene functions in a soil mobilome. Soil Swenson, N.G., 2014. Functional and Phylogenetic Ecology in R. Springer.

Biol. Biochem. 101, 175–183. Vryzas, Z., 2018. Pesticide fate in soil-sediment-water environment in relation to con-tamination preventing

actions. Curr. Opin. Environ. Sci. Health 4, 5–9. Mansour, I., Heppell, C.M., Ryo, M., Rillig, M.C., 2018. Application of the microbial community coalescence concept to riverine networks. Biol. Rev. Yuan, S., Tang, H., Xiao, Y., Xia, Y., Melching, C., Li, Z., 2019. Phosphorus contamination of the surface Masarikova, M., Manga, I., Cizek, A., Dolejska, M., Oravcova, V., Myskova, P., Karpiskova, R., Literak, I., sediment at a river confluence. J. Hydrol. 573, 568–580. 2003. Salmonella enterica resistant to antimicrobials in wastewater eﬄuents and black-headed gulls in the Czech Republic, 2012. Soil Biol. Biochem. 35 (4), 603–606. https://doi.org/10.1016/j.scitotenv.2015.10.069. Zolgharnein, H., Lila, M., Azmi, M., Saad, M.Z., Rahim, A., Abd, C., Mohamed, R., 2007. Detection of plasmids in heavy metal resistance bacteria isolated from the Persian Gulf and enclosed industrial areas, Iran. J. Monchy, S., Benotmane, M.A., Janssen, P., Vallaeys, T., Taghavi, S., Van Der Lelie, D., Mergeay, M., 2007. Biotechnol. 5, 232–239. Plasmids pMOL28 and pMOL30 of Cupriavidus metallidurans are specialized in the maximal viable response to heavy metals. J. Bacteriol. 189, 7417–7425.

137

Supporting Figure File Valentine Cyriaque, Samuel Jacquiod, Leise Riber, Waleed Abu Al-soud, David C. Gillan, Søren J. Sørensen, Ruddy Wattiez « Selection and propagation of IncP conjugative plasmids following long-term anthropogenic metal pollution in river sediments» Journal of Hazardous Materials (2019)

Figure S1: Map of Sensée canal (Férin, 50°18'39.0"N 3°05'05.4"E) and Deûle river (MetalEurop, 50°25'44.7"N 3°01'20.4"E) sampling sites. The figure was generated using Google Earth pro.

138

Figure S2: Log10 scatter plots of sorted particles obtained by flow cytometry for the selection of (i) bacterial cell through size scatter plots (left column; in blue, x-axis = front scatter FSC- A; y-axis = side scatter SSC-A); (ii) red fluorescent donor cells (middle column) and (iii) green fluorescent transconjugant cells (right column). Transconjugant cells were sorted based on their green fluorescence.

139

Figure S3: Plasmid transfer rate (T/(D*B)) obtained after conjugation assays and FACS, based on total number of cells sorted (B), transconjugant (GFP-expressing cells) and donor (mCherry-expressing cells). content of the bacterial community after filter mating assay in WWTP (Jacquiod et al., 2017) and in metal-impacted sediments (this study). Significance (0.001<**<0.01) was measured using a Kruskal-Wallis significance test.

140

Figure S4: Venn diagrams constructed with total community DNA (tcDNA) directly extracted from Férin’s and MetalEurop’s sediments by Jacquiod et al., 2018 as well as recipient and transconjugant taxonomic profiles obtained from in vitro conjugation assay using flow cytometry for sorting recipients (GFP expressing cells and unmarked cells) and transconjugants (GFP expressing cells).

141

Figure S5: Alpha-diversity profiles of recipient and transconjugant bacterial fractions (average ± SEM). Profiles were established from rarefied amplicon sequencing data (n=10.000) for the recipients and transconjugant (GFP expressing) cells sorted from Férin and MetalEurop samples. Letters indicates a significant difference between indices averages (ANOVA, Tukey’HSD, p<0.05).

142

Figure S6: Taxonomic distribution in Férin and MetalEurop total recipient communities and their transconjugant (GFP expressing cells) fractions obtained after filter mating assays on Férin and MetalEurop sediments.

143

Figure S7: Correlation between transfer rate (log transformed T/R) of each OTU and its activity in-situ (log transformed of total RNA/DNA) in Férin (A) and MetalEurop (B) SMCs. Pearson index of Férin SMCs was significantly positive. No significant correlation was measured in MetalEurop SMCs.

144

Supporting Table File Valentine Cyriaque, Samuel Jacquiod, Leise Riber, Waleed Abu Al-soud, David C. Gillan, Søren J. Sørensen, Ruddy Wattiez « Selection and propagation of IncP conjugative plasmids following long-term anthropogenic metal pollution in river sediments» Journal of Hazardous Materials (2019)

Table S1 : Total metal concentrations (mean ± SD; n = 4) in in Férin and MetalEurop sediments (0–1 cm) determined by Gillan and colleagues (2015) [1].

Férin MetalEurop Al(g/kg) 17.6±3.0 26.2±0.7 As(mg/kg) 2.8±0.3 21.0±0.9 Cd(mg/kg) 1.3±0.03 38.1±0.5 Co(mg/kg) 5.8±0.01 8.8±0.3 Cr(mg/kg) 56.2±1.7 107.4±1.9 Cu(mg/kg) 13.7±0.4 100.0±0.8 Fe(g/kg) 12.0±0.4 20.6±0.3 Mn(mg/kg) 293.5±7.6 547.9±1.4 Ni(mg/kg) 15.2±0.7 25.5±0.4 Pb(mg/kg) 111.6±0.8 913.8±11.0 V(mg/kg) 35.9±0.4 62.1±0.6 Zn(mg/kg) 348.5±6.7 3218.5±69.0

145

Table S2. Count table obtained from 16S rRNA amplicon sequencing Sort Station Counts Recipient Férin 21014 Recipient Férin 31944 Recipient Férin 20212 Recipient Férin 12700 Recipient Férin 18438 Recipient Férin 28293 Recipient Férin 35039 Recipient Férin 39480 Recipient Metal 37304 Recipient Metal 18338 Recipient Metal 29589 Recipient Metal 15272 Recipient Metal 30692 Recipient Metal 27965 Recipient Metal 25955 Recipient Metal 17968 Transconjugant Férin 33228 Transconjugant Férin 50709 Transconjugant Férin 31815 Transconjugant Férin 14892 Transconjugant Férin 29005 Transconjugant Férin 54223 Transconjugant Férin 26245 Transconjugant Férin 26994 Transconjugant Metal 21365 Transconjugant Metal 23283 Transconjugant Metal 27822 Transconjugant Metal 5087 Transconjugant Metal 39578 Transconjugant Metal 25484 Transconjugant Metal 22697 Transconjugant Metal 35778

Table S3 : Sequences of primers used for plasmid quantification in river sediments by qPCR.

Primer names Sequences Tm (°C) Reference Inc-P OriT Forwards - 5’-CAGCCTCGCAGAGCAGGAT-3′ 57 [2] Reverse - 5’-CAGCCGGGCAGGATAGGTGAAGT-3′ Inc-F OriT Forwards - 5′-TCTTCTTCAATCTTGGCGGA-3′ 52 [3] Reverse - 5′-GCTTATGTTGCACRGAAGGA-3′ Inc-I OriT Forwards - 5’-GGA GAG GAG ATC CGT TTC TGG-3’ 51 This study Reverse - 5’-TGC AGA CGG ATT TCA CTT TC-3’ Cf. Aknowledgement

146

Table S4: DNA template concentration and associated copy number results obtained by quantitative PCR. FER-DNA samples were extracted in the Férin sediments and MET-DNA samples were extracted in the MetalEurop sediments.

Plasmid Samples DNA templates (ng/µl) Number plamid copies per µg of DNA Incompatibility group IncF F_FER1 15,3 1,96E+03 IncF F_FER2 11,8 7,36E+03 IncF F_FER5 9,5 1,10E+04 IncF F_FER6 16,3 1,58E+03 IncF F_FER7 13,4 4,89E+03 IncF F_FER8 13,4 1,13E+03 IncI I_FER1 15,3 2,99E+03 IncI I_FER2 11,8 5,79E+03 IncI I_FER3 19,7 7,04E+02 IncI I_FER4 7,9 2,29E+03 IncI I_FER5 9,5 8,70E+03 IncI I_FER6 16,3 2,85E+03 IncI I_FER7 13,4 3,88E+03 IncI I_FER8 13,4 4,10E+03 IncP P_FER1 15,3 1,25E+06 IncP P_FER2 11,8 1,08E+06 IncP P_FER3 19,7 3,52E+05 IncP P_FER4 7,9 7,05E+05 IncP P_FER5 9,5 1,30E+06 IncP P_FER6 16,3 7,67E+05 IncP P_FER7 13,4 8,63E+05 IncP P_FER8 13,4 1,39E+06 IncF F_MET2 17,6 9,87E+02 IncF F_MET3 14,6 1,46E+03 IncF F_MET4 14,2 9,52E+03 IncF F_MET5 8,23 2,25E+03 IncF F_MET6 14,2 9,49E+02 IncF F_MET7 15,5 4,11E+03 IncF F_MET8 15,5 3,52E+03 IncI I_MET1 24 4,61E+02 IncI I_MET3 14,6 2,97E+03 IncI I_MET4 14,2 3,11E+03 IncI I_MET5 8,23 4,84E+03 IncI I_MET6 14,2 2,67E+03 IncI I_MET7 15,5 8,02E+03 IncI I_MET8 15,5 5,48E+03 IncP P_MET1 24 1,00E+06 IncP P_MET3 14,6 1,51E+06 IncP P_MET4 14,2 1,57E+06 IncP P_MET5 8,23 2,14E+06 IncP P_MET6 14,2 1,66E+06 IncP P_MET7 15,5 4,51E+06 IncP P_MET8 15,5 3,41E+06

147

Table S5. Beta-diversity assessed by rarefied (n = 10.000) abundance weighted and unweighted UNIFRAC OTU distance matrices using 1.000 permutations, carried out on recipients (GFP expressing cells and unmarked cells) and transconjugants (GFP expressing cells) after filter mating assays on Férin and MetalEurop sediments.

Factors tested r2 p-value Signif.

Weighted Unifrac CAP 1: Recipient/Tranconjugant sorting 0.76 9.9E-4 *** 2: Site (Férin/Metal) 0.12 4.0E-3 ** 1:2 0.07 0.02 * Unweighted Unifrac CAP 1: Recipient/Tranconjugant sorting 0.52 9.9E-4 *** 2: Site (Férin/Metal) 0.16 0.16 1:2 0.15 0.18

148

Table S6. Recipient from MetalEurop that do not uptake the plasmid

Sort OTU Phylum ( Proteobacteria sorted by class) Species Recipient from MetalEurop OTU_177 Betaproteobacteria Leptothrix_sp

Recipient from MetalEurop OTU_198 Epsilonproteobacteria Arcobacter_cryaerophilus

Recipient from MetalEurop OTU_1603 Firmicutes Sporosarcina

Recipient from MetalEurop OTU_129 Deltaproteobacteria Unclass_Desulfatitalea

Recipient from MetalEurop OTU_132 Bacteroidetes NLAE-zl-C268

Recipient from MetalEurop OTU_998 Firmicutes Paenibacillus_fonticola

Recipient from MetalEurop OTU_946 Firmicutes Paenibacillus_barcinonensis

Recipient from MetalEurop OTU_860 Firmicutes Bacillus_sp

Recipient from MetalEurop OTU_42 Firmicutes swine_fecal_bacterium_SD-Pec8

Recipient from MetalEurop OTU_174 Euryarchaeota Unclass_Methanosaeta

Recipient from MetalEurop OTU_850 Bacteroidetes Hymenobacter_fastidiosus

Recipient from MetalEurop OTU_1624 Actinobacteria Unclass_Conexibacter

Recipient from MetalEurop OTU_1251 Chloroflexi Unclass_Litorilinea

Recipient from MetalEurop OTU_763 Firmicutes Unclass_Firmicutes

Recipient from MetalEurop OTU_175 Alphaproteobacteria Brevundimonas_bullata

Recipient from MetalEurop OTU_855 Actinobacteria Unclass_Rhodococcus

Recipient from MetalEurop OTU_24 Gammaproteobacteria Acinetobacter

Recipient from MetalEurop OTU_472 Unclass_Proteobacteria Unclass_Proteobacteria

Recipient from MetalEurop OTU_636 Firmicutes Unclass_Staphylococcus

Recipient from MetalEurop OTU_52 Bacteroidetes Bacteroides_sp_enrichment_culture_clone_B3

Recipient from MetalEurop OTU_679 Gammaproteobacteria Morganella_morganii_subsp_null

Recipient from MetalEurop OTU_133 Firmicutes Unclass_Turicibacter

Recipient from MetalEurop & Férin OTU_384 Firmicutes Exiguobacterium

Recipient from MetalEurop & Férin OTU_78 Planctomycetes Unclass_Planctomycetaceae

Recipient from MetalEurop & Férin OTU_945 Firmicutes Bacillus_niacini

Recipient from MetalEurop & Férin OTU_662 Firmicutes Unclass_Bacillaceae

Recipient from MetalEurop & Férin OTU_166 Gammaproteobacteria Acinetobacter_soli

Recipient from MetalEurop & Férin OTU_571 Firmicutes Fictibacillus_arsenicus

Recipient from MetalEurop & Férin OTU_185 Firmicutes Paenibacillus_lautus

Recipient from MetalEurop & Férin OTU_358 Firmicutes Brevibacillus_sp_LY Recipient from MetalEurop & Férin OTU_61 Firmicutes Unclass_Solibacillus

Recipient from MetalEurop & Férin OTU_283 Firmicutes Kurthia_gibsonii

Recipient from Férin & MetalEurop and Férin Transconjugant OTU_464 Gammaproteobacteria Acinetobacter_sp_BN17

Recipient from Férin & MetalEurop and Férin Transconjugant OTU_1257 Gammaproteobacteria Haemophilus_parainfluenzae

Recipient from Férin & MetalEurop and Férin Transconjugant OTU_975 Alphaproteobacteria Unclass_Caulobacteraceae

Recipient from Férin & MetalEurop and Férin Transconjugant OTU_127 Actinobacteria Unclass_Ilumatobacter

Recipient from Férin & MetalEurop and Férin Transconjugant OTU_20 Firmicutes Lactobacillus_oligofermentans

Recipient from Férin & MetalEurop and Férin Transconjugant OTU_1499 Alphaproteobacteria Sphingomonas_adhaesiva

Recipient from Férin & MetalEurop and Férin Transconjugant OTU_68 Firmicutes Bacillus_subtilis

Recipient from Férin & MetalEurop and Férin Transconjugant OTU_315 Alphaproteobacteria Novosphingobium_resinovorum

Recipient from Férin & MetalEurop and Férin Transconjugant OTU_196 Betaproteobacteria Vogesella_indigofera

Recipient from Férin & MetalEurop and Férin Transconjugant OTU_827 Gammapoteobacteria Salmonella_enterica_subsp_enterica_serovar_Heidelberg_str

Recipient from Férin & MetalEurop and Férin Transconjugant OTU_922 Firmicutes Lysinibacillus_fusiformis

149

References

2. Malik A, Çelik EK, Bohn C, Böckelmann U, Knobel K, Grohmann E. Detection of conjugative plasmids and antibiotic resistance genes in anthropogenic soils from Germany and India. FEMS Microbiol Lett 2008; 279: 207–216.

3. Villa L, García-Fernández A, Fortini D, Carattoli A. Replicon sequence typing of IncF plasmids carrying virulence and resistance determinants. J Antimicrob Chemother 2010; 65: 2518–2529.

150

Chapter 4: Metal drives complex dynamics of IncP-plasmid in two-members bacterial community

Lead drives complex dynamics of conjugative plasmid

in bacterial community

Valentine Cyriaquea1, Jonas Stenløkke Madsenb, Laurence Fievezc, Baptiste Leroya, Lars Hansend, David C. Gillana, Fabrice Bureauc, Søren J. Sørensenb, Ruddy Wattieza a Proteomics and microbiology laboratory, Research Institute for Biosciences, University of Mons, 20 Place du Parc, Mons, Belgium b Section of Microbiology, Department of Biology, University of Copenhagen, Universitetsparken 15, 2100 Copenhagen Ø, 1, Bygning, 1-1-215, Denmark c Cellular and Molecular Immunology Service, Giga Research, University of Liège (ULiège), 11 Avenue de l’Hôpital, B-4000 Liège, Belgium d Environmental Microbial Genomics Group, Section for Environmental Microbiology and Biotechnology, Department of Environmental Science, Aarhus University, Roskilde, Denmark

1 Corresponding author: [email protected]; Tel.: +32 (0)65 37 33 19; Proteomics and microbiology laboratory, Research Institute for Biosciences, UMONS, 20 Place du Parc, Mons, Belgium

Abstract

Conjugative plasmids are vessels of accessory genes contributing to genome novelties, accelerating adaptation. Previously, plasmids have been shown to carry metal-resistance genes

(MRG) and were suggested to be ecological key players in the adaptation of metal-impacted microbial communities. However, direct metal impact on plasmid dispersion is far from being deciphered. In the present study, we used two-members bacterial communities to test the impact of lead on the dispersal of an IncP plasmid pKJK5 from a Pseudomonas putida KT2440 plasmid donor and two distinct recipients, Variovorax paradoxus B4 or Delftia acidovorans SPH-1.

Two versions of the plasmid were used, carrying, or not, the lead-resistance-associated pbrTRABCD operon in order to assess the importance of fitness benefit and conjugative potential for the dispersal of the plasmid. We shed light on specific conditions where lead promoted the dispersal of the broad-host range conjugative IncP plasmid in a recipient population. Dispersion of metal resistance conveyed by the conjugative plasmid required specific conditions of metal concentration and donor/recipient fitness. Furthermore, we showed modulating impact of lead on the conjugal-transfer machinery.

151

1. Introduction Metals constitute a serious risk for ecosystems because of their biotoxicity and bioaccumulation. They impact microbial communities by modifying their structure (1, 2), decreasing their diversity (2–4), and/or reducing activity rates of their members (5). Microbial communities were also shown to be able of impressive resilience and in several cases metals do not have long-term negative impacts on the diversity and activity of the communities (6, 7), insuring microbial functions for the ecosystem (8). Horizontal Gene Transfer (HGT), especially mediated by plasmid conjugation, is a key mechanism leading to the resilience of microbial communities (9, 10), in clinical and environmental ecosystems (11–13). They are vessels of genes contributing to genome innovation, making them key players for bacterial adaptation.

The persistence of conjugative plasmids in a community is dependent on the (i) rate of acquisition (transfer rate), (ii) fitness (i.e. ability to survive in a competitive environment) cost/benefits(iii) rate of loss (plasmid stability) (14, 15). Hence, if the burden of a plasmid is offset by a selection factor such as resistance to metals, or if conjugation efficiency is high enough, the persistence of the plasmid in the community is facilitated (15). Furthermore, fitness effects of the plasmid may also depend on specific genetic interactions between the plasmid and the host chromosome, as demonstrated across Pseudomonas species in a mercuric selective environment (16). Therefore, drawing a general scheme on the role of a metal as a selection factor for the maintenance and spread of plasmids in a microbial community is not trivial and requires the integration of multiple parameters. In Cupriavidus metallidurans CH34, metals increased the abundance of conjugative-transfer proteins (17) and cadmium has been shown to increase plasmid dispersal in subsurface-derived sediment microcosms (18). On the another hand, it has been shown that copper decreases conjugation frequencies in LB broth supplemented with copper nanoparticles (19). Furthermore, the overall impact of metals on plasmid dispersal in a soil microbial community was shown to be negative without impacting the diversity of transconjugants. Metals either negatively or positively modulated plasmid

152

uptake depending on the metal, metal concentration, and the evolutionary history of the recipients (20).

In the present work, we investigated the impact of lead (Pb (II)) on the transfer of a conjugative IncP-1ε plasmid, carrying, or not, the lead resistance operon pbrTRABCD, from

Cupriavidus metallidurans CH34. Transfer of pKJK5-gfp and pKJK5-gfp-pbr was investigated between a plasmid donor Pseudomonas putida KT2440 and a recipient strain, either Variovorax paradoxus B4 or Delftia acidovorans SPH-1. We assessed the impact of lead on plasmid dispersal in the recipient pool and the molecular adaptative response of the two-members communities.

2. Material and Methods 2.1. Strains, plasmids and growth conditions. For the construction of plasmid pKJK5-gfpmut3-pbrTRABCD-kanR-tetR, Cupriavidus metallidurans CH34 and Electrocomp™ GeneHogs® E. coli (Invitrogen) were cultivated in LB medium (see section 2.2. for details on how pbrTRABCD operon was inserted into the plasmid).

R The strain Pseudomonas putida KT2440::PlppmCherry-kan harbouring the plasmid pKJK5- gfpmut3-kanR-tetR (pKJK5-gfp) or pKJK5-gfpmut3-pbrTRABCD-kanR-tetR (pKJK5-gfp-pbr, see section 2.2) was used as plasmid donor. Precultures of those strains were grown in 3-time diluted LB broth buffered with MOPS (2.1 g/L) (LB3D) supplemented with 50 µg/mL tetracycline (30°C) overnight. Variovorax paradoxus B4 (DSMZ, Germany) and Delftia acidovorans SPH-1 (DSMZ, Germany) were used as plasmid recipient. Precultures of these strains were grown overnight in 457 medium (DSMZ) supplemented with 2 g/L mercaptosuccinate and LB3D, respectively (30°C). For the plasmid burden assay, both plasmid donors and the clones V. paradoxus B4:: pKJK5-gfp, V. paradoxus B4 - pKJK5-gfp-pbr, D. acidovorans SPH-1:: pKJK5-gfp and D. acidovorans SPH-1::pKJK5-gfp-pbr were obtained by conjugation between those recipient cells and either GeneHogs® E. coli- pKJK5-gfp or

GeneHogs® E. coli- pKJK5-gfp-pbr (Supporting method). Plasmid burden assays and

153

conjugation assays were performed in 15 mL LB3D supplemented with 0, 0.5, 1 or 1.5 mM of

Pb (NO3)2 at 16°C and 120 RPM in triplicate (see section 2.3 and 2.4). Plasmid construction and transformation of pKJK5-gfp-pbr is described in Supporting method.

2.2. Assessing plasmid burden The burden of plasmids pKJK5-gfp and pKJK5-gfp-pbr when carried by Pseudomonas

R putida KT2440::PlppmCherry-kan , Variovorax paradoxus B4 or Delftia acidovorans SPH-1 was measured relatively to corresponding plasmid free strains as the ratio between the growth rate of plasmid free and the plasmid carrying cells (µ0/µ1) (15) considering that pKJK5 displays a low rate of plasmid loss (21). Plasmid carrying clones were obtained by conjugation as described in section 2.1. and selected on tetracycline. The strains were pre-cultured in 100 mL

LB3D (30°C) (supplemented with tetracycline 50 µg/mL for plasmid carrying strains) until the

Optical Density (OD, 600 nm) was between 0.4 and 0.6 measured in cuvettes using a Helios

Zeta UV-Vis (Thermo Scientific). Then, precultures were washed twice in LB3D (2 minutes,

7000×g) and cells were counted using Bright-Line™ Hemacytometer (Merck) following manufacturer instructions. A stated number of cells (Table S1) were sampled from precultures and diluted in a final volume of 15 mL LB3D supplemented with 0; 0.5; 1 or 1.5 mM of Pb

(NO3)2.in 50 mL falcon (n=3). The growth rates of plasmid free and plasmid carrying strains were obtained by growing them (16°C, 120 RPM) until the lag phase. OD (595 nm) was measured on 96-wells multiplate with FLUO star OPTIMA (BMG LabTech). Significance of the burden for each strain at different lead concentrations was assessed by comparing their growth rate growth with plasmid-freed corresponding strain.

2.3. Conjugation assay in liquid community and cytometry analysis Plasmid transfer rates of plasmids pKJK5-gfp and pKJK5-gfp-pbr carried by

R Pseudomonas putida KT2440::PlppmCherry-kan as plasmid donor was assessed in co-cultured with Variovorax paradoxus B4 or Delftia acidovorans SPH-1 (Figure 1) with the equation

T/(DR) where “D” is the count of donors (plasmid carrying Pseudomonas putida

154

R KT2440::PlppmCherry-kan ); “R” is the count of recipients (plasmid free V. paradoxus B4 or

D.acidovorans SPH-1 cells) and “T” is the count transconjugants (recipient cells that acquired the plasmid). To start co-cultures, the strains were pre-cultured in 100 mL LB3D (30°C)

(supplemented with tetracycline 50 µg/mL for plasmid carrying strains) until OD (600 nm) was between 0.4 and 0.6. OD was measured in cuvettes using a Helios Zeta UV-Vis (Thermo

Scientific). Then, precultures were washed twice in LB3D (2 minutes, 7000×g) and cells were counted using Bright-Line™ Hemacytometer (Merck) following manufacturer instructions. A stated number of cells (Table S1) were sampled from preculture and diluted in a final volume of 15 mL LB3D supplemented with 0; 0.5; 1 or 1.5 mM of Pb (NO3)2 in 50 mL falcon (n=3).

Co-cultures were grown at 16°C, 120 RPM for 10 days and sampled (i) after 4 days for cell count and proteomics analyses (see section 2.5) and (ii) after10 days for cell count (Figure 1).

Donor, recipient and transconjugant counts were assessed by flow cytometry for the detection

GFP fluorescence expressed by the gfpmut3 gene carried by the pKJK5 plasmid carried by donor and transconjugant cells. Plasmid donor cells were differentiated from transconjugants by Fluorescence In-Situ Hybridization (FISH) targeting the 16S rDNA gene of Pseudomonas putida using a modified protocol from Gougoulias and Shaw (2012) (22). Briefly, 1 mL of co- culture was centrifugated (6 minutes, 6000×g) and fixed by resuspending it in 1 mL PFA 4%; pH7 for 15 minutes. Fixed cells were washed twice in PBS (6 minutes, 6000×g) and resuspended in 247 µl of prewarmed (48◦C) hybridization buffer added with 3 µL of the probe

PSE1284 (23) associated to a AlexaFluor647 fluorochrome (Eurogentec, Liège, Belgium) (250 ng/µL) and incubated for 4 hours. Samples were centrifugated (5 min, 16,000×g) and resuspended in 500 µL of hybridization buffer for 20 minutes. They were centrifugated (5 min,

16,000×g) again and resuspended in 500 µL of wash buffer. Finally, samples were centrifugated, resuspended in 1 mL PBS and stored at 4°C until cytometry analyses. As hybridization and wash buffers, urea solutions were used to avoid GFP denaturation

155

(Hybridization buffer: urea 1M, NaCl 0.9M, Tris HCl pH 7.4 20 µM; Wash buffer: urea 4M,

NaCl 0.9M, Tris HCl pH 7.4 20 µM) as described before (24, 25).

Cytometry analyses were carried out using BD Science FACS Fortessa in 96-wells microplate with the following parameters: forward scatter 390 V, side scatter 176 V, detectors for green fluorescence associated to GFPmut3 fluorescence (bandpass filter 530/30 nm, 501 V) and for red fluorescence associated to AlexaFluor 647 (bandpass filter 670/14 nm, 550 V).

FlowJo V10 was used for analysing results (e.g. Figure S1) to count donor, empty donor, recipient and transconjugant (Table S2).

2.4. Proteomic analysis by SWATH mass spectrometry A quantitative proteomic approach was used to assess metal impact on plasmid transfer machinery and interactions in bacterial consortia. For that, 600 µL of 0 mM and 0.5 mM of lead cultured samples and 1200 µL of 1 mM and 1.5 mM of lead cultured samples were harvested, centrifuged (6,000×g, 6 min, 4°C) and washed twice (6,000×g, 6 min, 4°C) with PBS. Proteins were extracted, reduced, alkylated, precipitated and trypsinized from the pellet using the

PreOmics kit (PreOmics GmbH, Planegg, Germany) following the manufacturer instructions.

Obtained peptides were quantified using Pierce™ Quantitative Colorimetric Peptide Assay

(ThermoFisher Scientific). For post-acquisition retention time calibration, a PepCalMix solution (Protein Extract Digest) (AB SCIEX, Framingham, MA) was added to 4 µg of peptides

(50 fmol on column) following manufacturer instructions. Peptides (2 µg on column) were analysed on a UHPLC-HRMS/MS instrument (AB Sciex LC420 & TripleTOF™ 6600) using

SWATH data-independent acquisition (DIA) as described in Supporting method. SWATH wiff files were processed using the AB Sciex PeakView 2.2 software and his SWATH™ Acquisition

MicroApp. Up to 6 peptides with at least 95% confidence were selected with 6 transitions per peptide. XIC extraction window was set to 15 min, and XIC width to 70 ppm. The peptide area corresponds to the sum of the fragment ion area and the protein area corresponds to the sum of

156

peptide area. Protein area were extracted and exported in AB Sciex MarkerView™ 1.2 software for normalization and statistical analysis. Protein extraction characteristics, the number of proteins identified in both libraries at 1% FDR as well as the proteome coverage of the SWATH proteomic analyses are displayed in tables S3 and S4. Protein abundances were normalized by the cumulated protein area. The obtained dataset underwent different normalizations.

Abundances of chromosomal encoded proteins were either normalized by the proportion of the respective specie in the two-member communities obtained by flow cytometry (see section 2.4.) or normalized by the cumulated protein abundance of proteins assigned to the respective strain.

Plasmid encoded proteins were normalized by proportion of plasmid carrying cells in the two- members community. Results were deposited on Peptide Atlas public repository

(http://www.peptideatlas.org/) under the accession number PASS01468

(http://www.peptideatlas.org/PASS/PASS01468). Significant differences between norm2- protein abundances (log-2 transformed) were determined using a two-tailed Student’s t-test across the different lead concentrations for each filter mating association and for the comparison of same lead condition impacting the mating pair with or without pbrTRABCD in the exchanged plasmid. Proteins displaying a minimum p-value below 0,05 were taken into consideration and strain-specific proteins with a p-value below 0,01 were plotted in heatmaps using hierarchical cluster dendrograms (Euclidean distance and average clustering) on the RGui software (vegan

(26), rioja (27) and gplots (28) R-packages). Strain-specific proteins are displayed in Heatmap supporting information.

157

Figure 1: Experimental design gathering flow cytometry and proteomics analysis to assess metal impact on the functional profile of a mating co-culture between a pKJK5 plasmid and link it to the plasmid transfer rate. Recipient and donor cells were grown (10/1 ratio) in LB3D medium (120 RMP, 16°C) during 10 days with an increasing concentration of Pb2(NO3)2 (0, 0.5, 1, 1.5 mM). After 4 and 10 days, they were analysed by flow cytometry to quantify recipient, donor (FISH labelling) and transconjugant cells. After 4 days, they were also analysed by SWATH-MS Proteomics. MS: Mass Spectrometry; m/z: mass-to-charge ratio; XIC: Extracted ion chromatogram.

158

2.5. MRM identification and quantification of the PbrA protein Quantification of PbrA was performed using Multiple Reaction Monitoring (MRM) based relative quantification. Spectral signature of PbrA was obtained through the analysis of

Cupriavidus metallidurans CH34 samples using regular LC-MS/MS procedure (29). This sample was selected because C. metallidurans is known to express, in presence of high lead load, high level of PbrA. Bacterial culture was obtained as described previously (30) and proteins were extracted and digested using PreOmics kit (PreOmics GmbH,

Planegg/Martinsried) and peptides quantified using Pierce™ Quantitative Colorimetric Peptide

Assay (ThermoFisher Scientific), following manufacturer instructions. A total of 13 peptides were detected in LC-MS/MS analysis of C. metallidurans and evaluated for quantification in

MRM mode on a Q-Trap 6500+ coupled to a LC420 chromatography (Sciex) operated in microflow mode. After transition optimisation and interference removal, 4 peptides with 4 to 5 transitions each were selected for quantification (Table S5). Quantification was assessed on (i) pKJK5-gfp carrying sample at 1 mM of lead as control as (ii) all pKJK5-gfp-pbr carrying samples. For that,3µgr of trypsin digested proteins were separated on a C18 YMC-Triat 0.3x150 mm column operated at a flow rate of 5 µl/min with an ACN gradient from 2 to 35% in 17 minutes. Skyline was used for visual inspection of the data and area under the curve integration.

Peak picking for each peptide was manually refined using transition intensity ratio and retention time as leading parameters. These parameters were obtained from C. metallidurans sample which contained higher level of PbrA. Intensity of all 4 to 5 transitions were summed for each peptide. Protein abundance was obtained as the average of Ln transformed area under the curve of each four peptides.

3. Results 3.1. Plasmid burden is different across mating species The burden of the pKJK5-gfp and pKJK5-gfp-pbr plasmids impeding the growth of plasmid donor and recipients was calculated as the ratio of the growth rates of plasmid free and plasmid

159

carrying cells. Despite a decreasing trend in the growth rate of V. paradoxus B4, compared to plasmid free cells, no significant variation was detected. Consequently, plasmids had no significant burden for V. paradoxus B4 (Figure 2). Growth rates of D. acidovorans SPH-1 were significantly impacted at high lead concentration (1 and 1.5 mM) for pKJK5-gfp carrying cells, imposing a burden to the cells. This burden is apparently compensated by the pbrTRABCD operon, dedicated to lead resistance, since growth rate of the strain carrying these resistance genes was equivalent to WT at high concentration of lead. Pb (1 and 1.5 mM) had negative impact on the growth rate of plasmid carriers P. putida KT2440. Nevertheless, at 1 mM lead, only cells carrying pKJK5-gfp were negatively impacted. At 1.5 mM of lead concentration, decreased growth was significant in both plasmid carriers with subdued effect on pbrTRABCD carrying cells as confirmed by the calculated burden (Figure 2).

Figure 2: Growth rates (µ) and plasmid burden (µ0/µ1) ±SEM where µ0= growth rate of the plasmid- free cells and µ1= growth rate of plasmid-carrying cells; in pure cultures. Pp: Pseudomonas putida KT2440; Vp: Variovorax paradoxus B4; Da: Delftia acidovorans SPH-1; G: pKJK5-gfpmut3-kanR-tetR; GP: pKJK5-gfpmut3-pbrTRABCD- kanR-tetR. Stars show significant differences among growth rates and the significance of the burden (comparison of growth rate of the plasmid-carrying cells and plasmid- free cells at the same lead concentration. *: p-value<0,05; **:p-value<0,01; ***:p-value<0,001 (N=3) (n=3; Tukey test; p≤0.05).

160

3.2. Two-members community dynamics assessed by flow-cytometry To apprehend the dynamic of the two-members communities over time, we used a flow- cytometry approach to measure the proportion of the different strains and plasmid-carrying cells, (donor P. putida KT2440 and transconjugant cells).

First, when co-cultured with D. acidovorans SPH-1, the proportion of P. putida KT2440 was lower than in the co-culture with V. paradoxus B4 (Table S2), revealing a high relative fitness of Delftia acidovorans SPH-1. Besides, the plasmid transfer rate per recipient cells was higher when V. paradoxus B4 was used as recipient (Figure 3).

As shown in Figure 3, lead only had a significant negative impact on plasmid transfer rate of pKJK5-gfp into recipient Variovorax paradoxus B4 at the concentration of 1 mM, both after

4 and 10 days. When D. acidovorans SPH-1 was used as recipient, Pb overall had a negative impact on the transfer rate. Interestingly, (i) at 1 mM of lead, the transfer rate of plasmid pKJK5- gfp-pbr was not impacted and (ii) at 1.5 mM, an increased transfer rate of pKJK5-gfp plasmid was recorded after 4 days, but transconjugants relapsed after 10 days.

When Variovorax paradoxus B4 was used as recipient, plasmid loss in the donor cell fraction (plasmid free P.p.; Table S2) progressively increased with lead concentrations and with time. When D. acidovorans SPH-1 was used as recipient, plasmid loss in the donor P. putida

KT2440 was high at 1.5 mM of lead (84,3±4,2 % of cured cells from pKJK5-gfp; 74,1±1,4 % from pKJK5-gfp-pbr) after 4 days of co-culture. After 10 days, plasmid free P. putida KT2440 largely decreased (18,5±12,6 % of cured cells from pKJK5-gfp; 11,4±3,1 % from pKJK5-gfp- pbr) (Table S2).

161

Figure 3: Plasmid transfer rate (T/(D*R)) ±SEM into the recipient cells, Variovorax paradoxus B4 (A, C) or Delftia acidovorans SPH-1 (B, D) after 4 days (A, B) or 10 days (C, D) of mating. p-values were calculated from log-2 transformed abundances using t-test (4-days mating, black stars) or Test de Kruskal-Wallis test (10 days-mating grey stars) depending on homoscedasticity. *: p-value<0,05; **:p- value<0,01; ***:p-value<0,001 (N=3). Dotted lines are for pKJK5-gfp comparisons and continuous lines are for pKJK5-gfp-pbr comparisons.

3.3. Meta-proteomic profiling of two-members communities To gain better understanding into impacts of lead on two-members communities and their conjugation mating, depending on the nature of the plasmid, a quantitative meta-proteomic analysis was performed using a SWATH approach. For this, a spectral library built with a Data

Dependent Acquisition (DDA) workflow was generated with the three strain monocultures and in co-culture with P. putida KT2440 (Table S1) at 0 mM and 1 mM of Pb (NO3)2.

First, the label free DDA analyses of pure cultures grown in the presence or not of 1 mM lead revealed the induction of metal-resistance-associated proteins (Table S6). In Pseudomonas putida KT2440, the induction of phosphate metabolism-associated proteins (e.g. phosphatase)

162

and pyoverdine-associated proteins was observed. The ATPase PbrA was exclusively observed in the presence of lead when the operon was carried on the pKJK5 plasmid. In Variovorax paradoxus B4, a high level of TonB siderophore-associated proteins were induced as well as a putative ABC transporter iron-binding protein, and the iron-sulfur assembly scaffold protein

IscU. In Delftia acidovorans SPH-1, a large diversity of metal-resistance-associated proteins were observed: (i) phosphate metabolism-associated proteins, (ii) TonB siderophore-associated proteins, (iii) efflux pumps, (iv) a glutathione S-transferase, (v) the iron-sulfur assembly protein

IscA, (vi) a thioesterase, (vii) an iron permease and (viii) bacterioferritin (Table S6).

From SWATH-MS results, strain abundances can be evaluated as the sum of area of all proteins attributed to the specific strain. Nevertheless, flow cytometry and proteomics analyses revealed disparities in term of the relative abundance of the strains in each tested two-members communities (table S2). It is well known that a large expression of proteins associated to a specific strain - for example if in a two-members community, one strain is more metabolically active than the other - leads to a potential misestimation of the relative abundance of this strain in the bacteria community (31). Furthermore, we showed that this ratio between cell and protein counts was particularly unbalanced at 1.5 mM of lead, especially in co-culture of P. putida

KT2440 and D. acidovorans SPH-1 (Figure 4, Figures S2-S5). In this context, protein fold- changes for each strain between the control (0 mM Pb) and different lead concentrations was determined after a data-normalization based on the total protein abundance associated to the specific strain.

163

Figure 4: Patterns of differential abundance of P. putida KT2440’s protein (when associated to D. acidovorans SPH-1) were displayed by volcano plot according to the log2 of the fold- change ratios and the negative log10 of the p-value. Fold change and p-value (t-test on log2 abundance) were calculated for each proteins ( ≥2 peptide identifications) after normalisation on summed area of all proteins for each sample (A), and normalization on cell ratio in the two- members community obtained by flow cytometry (B) or with normalization on summed area of all proteins attributed to the specific strain ( i.e. Pseudomonas putida KT2440:: pKJK5- R R gfpmut3-kan -tet ) (C). Dot lines demarcates thresholds of fold changes below 0,66 (log2(0,66)

=- 0.58) and over 1.5 (log2(0,66) = 0.58) and p-value below 0,05 (-log10(0,05) =1,30). Entire set of volcano plots is displayed in figures S2-S5.

3.3.1. Metabolic impacts and resistance mechanisms In addition to unbalanced cell/protein ratio, most underabundant P. putida KT2440’s proteins were associated to main metabolism of the cell relating a decrease of bacterial abundance (Heatmap supporting information including Figure S6). When the strain was co- cultured with D. acidovorans SPH-1, this decrease was particularly high and was reduced when the pbrTRABCD operon was on the plasmid, revealing its benefit for the plasmid donor in that mating interaction. Detailed metabolic differences in the 3 strains’ proteomes in both mating- pair conditions with increasing lead concentrations are displayed in Heatmap supporting information. Most apparent lead-impacts concern enzymes depending on divalent cation co-

164

factors (Fe (II), Mg (II), Zn (II), Co (II), Mn (II)) or on the [4Fe-4S] cluster at 1.5 mM.

Interestingly, in D. acidovorans SPH-1, Mg (II)-dependant enzymes were over-abundant as well as Mg (II) import-associated proteins while Pb (II) had negative impact on its mating partner P. putida KT2440’s enzymes using Mg (II). Lead up-regulated Mg (II) import- associated proteins of this strain suggesting competition for this essential element between the two. Furthermore, in parallel to decreased abundance of the [4Fe-4S]-dependant enzyme fumarate hydratase class I, all strains resorted to the iron-independent fumarate hydratase class

II (Figure 5C). Indeed, despite lead cannot directly inactivate the [4Fe-4S]-dependant Class I enzyme (32), the oxidative stress that lead most likely induced (33) decreased the abundance of this enzyme as previously shown with Al-Ga toxicity (34). In P. putida KT2440, this last enzyme was much less abundant when grown with D. acidovorans SPH-1 highlighting the decreased fitness of the donor strain when facing D. acidovorans SPH-1. Similarly, the decreased abundance of the succinate dehydrogenase iron-sulfur subunit of P. putida KT2440 was amplified in the presence of D. acidovorans SPH-1 (Figure 5 A & B) and reduced with the benefit of the pbrTRABCD operon. In V. paradoxus B4, an increased amount of pilus related proteins (Heatmap supporting information) may explain the high transfer frequency into this recipient strain, compared to transfer rates in D. acidovorans SPH-1 (Figure 3).

At high lead concentration, all strains up-regulated metal-resistance-associated proteins involving phosphates (production of phosphate salt for the precipitation of metal cations (35)), metabolism of glutathione (bind metal cations (30)) and efflux transporters (RND efflux systems and heavy-metal transporter) (P COG) (Heatmap supporting information). In V. paradoxus B4’s proteome, additional metal-resistance involved proteins were also up- regulated, such as TRAP dicarboxylate transporters, TonB receptors (P COG), a putative non- ribosomal peptide synthase (NRPS) and the polyketide synthase(Q COG), for the biosynthesis of variochelin lipopeptide siderophore (36). In D. acidovorans SPH-1, a large arsenal of metal-

165

resistance-induced proteins were over-abundant, including TonB siderophores, phosphatases, and efflux pumps, explaining its high fitness (P COG) (Heatmap supporting information). P. putida KT2440 also up-regulated pyoverdine-associated proteins (V COG) for the binding of

Pb(II) (30, 35, 37). When co-cultured with D. acidovorans SPH-1, P. putida KT2440 up- regulated additional proteins associated to metal-resistance functions, including, a PHA storage-related protein (I COG) and the SurA chaperone (O COG), associated to the assembly of outer membrane proteins. (Heatmap supporting information). Also, the penicillin binding protein 1B was more increased at 1.5 mM of lead with a lower increase when P. putida KT2440 carry lead resistance operon pbrTRABCD (Figure 5B). At 1.5 mM of lead, when pbrTRABCD operon was encoded on the pKJK5 plasmid, the toluene efflux periplasmic linker protein TtgA was increased as well as the outer membrane protein H1 (Figure 5B), decreasing the absorption of cations absorbed on lipoproteins (38). At 1 mM of lead, phosphatase, phosphate-binding and transport-associated proteins as well as putative cation transporter (P COG) and PhoB and PhoR proteins (T COG) were more abundant when pbrTRABCD operon was present (Figure 6).

166

Figure 5: Abundances (±SEM) of proteins of interest in two-members community without the pbrTRABCD operon (G) or including the pbrTRABCD operon (GP). The number indicates the Pb (II) concentration (mM). Different panels represent abundance of proteins identified as belonging to P. putida KT2440 when cultivated with Variovorax paradoxus B4 (A) or with D. acidovorans SPH-1(B); and the abundance of fumarate hydratase proteins in P. putida KT2440 when cultivated with Variovorax paradoxus B4 (C1); in Variovorax paradoxus B4 (C2); in P. putida KT2440 when cultivated with D. acidovorans SPH-1(C3) and in D. acidovorans SPH-1 (C4). p-values were calculated from log-2 transformed abundances using t-test (n=3). *: p-value<0,05; **: p-value<0,01; ***: p-value<0,001.

167

Figure 6: Centred-scaled log-2 transformed abundances of P. putida KT2440’s (P.p.) and D. acidovorans SPH-1’s (D.a.) proteins (≥2 identified peptides) of interest belonging to P (Inorganic ion transport and metabolism) and T (Signal transduction mechanisms) COGs displayed in heatmaps (Euclidean distance, average clustering; entire heatmaps: see Heatmap supporting information).

3.3.2. Impacts on the pKJK5 backbone genes To decipher lead effects on the conjugation machinery, we also looked at impacted conjugation-associated proteins encoded by the pKJK5 plasmids. We normalized the abundance of these proteins by the proportion of plasmid carrying cells assessed by flow cytometry. When V. paradoxus B4 was the mating partner, 16 proteins encoded by the pKJK5 plasmid were significantly impacted. Among them, TraD, TrbA, TrbB, TrbF, TrbG and TrbH, were decreased at 1.5 mM (Figure 7). When D. acidovorans SPH-1 was the mating partner, 15 proteins encoded by the pKJK5 plasmid were significantly impacted. At the community level

(no cell normalization, Figure S7), when the pbrTRABCD operon is not carried by the pKJK5 plasmid, 10 proteins for the conjugative transfer (KorA, KorB, InC1, TraD, TraE, TraL, TrbE,

TrbF, TrbH, TrbM,) were significantly underabundant at the Pb 1.5 mM (Table S7). When the pbr operon was on the plasmid, this decrease was not significant for most proteins excepted for

KorA, TraG and TraL. At 0.5 and 1 mM of lead, conjugation-associated proteins were increased

168

only if pbr operon was carried on the plasmid, compared to the control (0 mM Pb). At the cell level (additional normalization by the plasmid carrier proportion), proteins KorB, IncC1, TraC,

TraE, TraG, TrbE and TrbI were increased at 1.5 mM of lead, especially when the pbrTRABCD operon was not on the pKJK5 plasmid (Figure 7, Table S7). At 1 mM of lead, the fold change was particularly high for TraC and TraG proteins (Figures 5 and S7, table S7).

Figure 7: Abundance of conjugation-associated proteins (≥2 identified peptides) encoded by the pKJK5 plasmid (normalisation on summed area of all peptides for each sample and by cell proportion of plasmid carrier (donor + transconjugant) (see Table S7)) displayed on a heatmap (Euclidean distance and average clustering) as extracted from two-members communities using either Variovorax paradoxus B4 (A) or D. acidovorans SPH-1 (B) as plasmid recipient.

169

3.3.3. Plasmid encoded pbrA expression Pbr proteins were not detected in the SWATH analysis revealing their low abundance level. Therefore, PbrA relative abundance was measured for the samples carrying the pKJK5- gfp-pbr plasmid using a targeted MRM-based relative quantification approach (Figure S8). This very sensitive method revealed the presence of PbrA in samples. No significant differences in the Variovorax paradoxus B4 mating pair, and an increased relative abundance in the D. acidovorans SPH-1 mating pair (at Pb 1 and 1.5 mM) were noticed (results were normalized by the proportion of plasmid carrier cells, as determined by flow cytometry; i.e. donors and transconjugants).

4. Discussion Previous works revealed that metals either favoured (17, 18) or inhibited bacterial conjugation (19, 20) with differential impact on the bacterial members of a soil community

(20). It was suggested that the burden and advantage trade-offs imposed by plasmids make their dispersion a dynamic process (10, 39). The pKJK5 plasmid has been shown to insure its stability in a bacterial population using a high conjugation efficiency (40). Then, lead may either modulate its conjugation efficiency or topple the burden/benefit ratio.

4.1. Lead, mating partner and beneficial genes modulated the fitness of the hosts Separately, P. putida KT2440, V. paradoxus B4 and D. acidovorans SPH-1 displayed similar growth rates and similar lead Minimal Inhibitory Concentrations (MIC, 2 mM in LB

3D), however, the lead-resistance mechanisms they induced were different. Variovorax is known for its potential for siderophore formation (36) and indeed mainly up-regulated siderophore-associated proteins (e.g. NRPS, polyketide synthase and outer-membrane TonB receptor). Furthermore, the burden of the pKJK5 plasmid on V. paradoxus B4 was not increased by lead neither by the presence of the pbr operon. PbrA is an efflux pump transporting zinc, cadmium and lead ions from the cytoplasm to the periplasm where phosphates, released in the periplasm by the undecaprenyl pyrophosphate phosphatase PbrB, form metallo-phosphates

170

(41). The external metal-binding function of siderophores may have made the periplasmic phosphatase PbrB useless. Consequently, the pbr operon brought no benefits for V. paradoxus

B4 (41) and the plasmid transfer efficiency was only dependant on the fitness of the donor. At

Pb 1mM, the dispersion of the pKJK5-gfp plasmid was significantly decreased, concurrently to its increased burden in P. putida KT2440. At 1.5 mM, the decrease of transfer efficiency was not significant. However, following to the decreased abundance of conjugation-associated proteins, plasmid loss in the plasmid donor increased, decreasing the rate of donor cells. When the plasmid brought no beneficial trait to the recipient, lead had an overall negative impact on the spread of the plasmid.

Noteworthy, when cultured in two-members communities, the total abundance of proteins from P. putida KT2440 cells was clearly decreased at 1.5 mM of lead, in both communities.

This is in agreement with previous findings that cell size, and hence the protein abundance, was decreased in the presence of metals in Pseudomonas aeruginosa 4EA (37). A decreased protein abundance may also result in a decreased activity of this bacterium. Furthermore, both when assessed using flow cytometry (cell counts) and metaproteomics (specific protein counts), the relative fitness of P. putida KT2440 was decreased when grown with D. acidovorans SPH-1.

In this mating couple, P. putida KT2440’s protein abundance decreased at high metal- concentration but D. acidovorans SPH-1’s protein abundance increased. The competitive interaction between P. putida KT2440 and D. acidovorans SPH-1 was also shown by the magnesium metabolism of both strains, the moderate up-regulation of the [4Fe-4S]-independent

Class II fumarate hydratase by Pseudomonas putida KT2440 and the negative impact of Pb 1.5 mM on Pseudomonas putida KT2440’s succinate dehydrogenase. The impact of 1.5 mM lead was subdued by the presence of the pbr operon, revealing the fitness it procures, especially in a context of competition. Furthermore, the burden of the plasmid was ameliorated by the presence of the pbr operon at 1 mM of Pb leading to a null burden, as assessed by increased

171

abundance of PbrA. The specific lead efflux/sequestration system encoded by pbrA and pbrBC

(30) may have enabled alternative systems such as the outer membrane protein H1, the toluene efflux pump TtgA, phosphatases and proteins associated to phosphate-import, especially at Pb

1 mM. Likewise, the phosphate-import-related protein over-abundance at that concentration may then be required to insure the turnover of undecaprenyl pyrophosphate in the membrane.

This differential meta-proteomic profile between pKJK5-gfp and pKJK5-gfp-pbr carrying cells reveal the ability of the hosts to capitalize on the Metal-Resistance Genes (MRGs). Beneficial genes then allowed to maintain the speediness of dispersion of the plasmid after 4 days of mating (compared to Pb 0 mM) and to contain it in the recipient population over 10 days.

4.2. Lead impact on the partitioning system and conjugative machinery. When mating V. paradoxus B4 and P. putida KT2440, plasmid cured cells increased at 1.5 mM of lead over 10 days, concurrently to a down-regulation of the conjugative machinery.

When mating D. acidovorans SPH-1 and P. putida KT2440, plasmid cured cells increased at 1.5 mM of lead after 4 days. However, conjugation-involved proteins (TraG, TraC, TrbE,

IncC1 and KorB) were up-regulated at that concentration, especially from the pKJK5-gfp plasmid, resulting in (i) the recovery of the donor population after 10 days, most probably helped by IncC1 partitioning protein, and (ii) a fast dispersion of the pKJK5 plasmid after 4 days of mating. The conjugative coupling protein (T4CP, TraG) and the DNA primase TraC allowing plasmid replication in the transconjugant (42–44), ensured the fertility of conjugation

(42). At the community level, the abundance of conjugation-associated proteins was decreased at 1.5 mM of Pb, leading to a decreased dispersion on a longer term (10 days) where plasmid settlement was most likely determined by its burden. These results confirm that, in some conditions, metals may promote the expression of conjugative transfer proteins, as shown in C. metallidurans CH34 (17), while fine regulation of the machinery must be deciphered.

172

To conclude, as metals, specifically lead, influence many cellular processes and mobilise different resistance strategies, MRGs encoded on the plasmid provide a limited benefit that depend on the host cell and metal concentration. MRGs conveyed by conjugative plasmids into a recipient cell population then require specific conditions of metal concentration and donor/recipient fitness. Plasmid dispersion depend on the ability of the recipient cell to capitalize on these MRGs, then, selecting for MRG-carrying plasmids at specific lead concentration. Interestingly, it was shown that, at high lead concentrations, when the fitness of the recipient D. acidovorans SPH-1 was high, the conjugal-transfer machinery was up- regulated. The present study unravelled the dynamics of propagation of a broad-host-range conjugative IncP plasmid and its role in the spread of MRGs in a lead-enriched environment.

These results suggest that metals may influence the journey of plasmids in a diversified recipient community leading to the enrichment of MRGs at the community level.

5. Acknowledgments This study was sponsored by the “Belgian Fund for Scientific Research” (Grand equipment-F.R.S.-FNRS), by the Pole d’ Attraction Inter-universitaire (PAI) n◦ P7/25 and by the Fund for Scientific Research (F.R.S-FNRS) FRFC 7050357, PF014F828/A09016F and

T.0127. The Bioprofiling platform used for the proteomic analysis was supported by the

European Regional Development Fund and the Walloon Region, Belgium.

6. References 1. Gillan DC, Danis B, Pernet P, Joly G, Dubois P. 2005. Structure of Sediment-Associated

Microbial Communities along a Heavy-Metal Contamination Gradient in the Marine

Environment. Appl Environ Microbiol 71:679–690.

2. Guo Q, Li N, Xie S. 2019. Heavy metal spill influences bacterial communities in freshwater

sediments. Arch Microbiol 201:847–854.

3. Kwon MJ, Yang JS, Lee S, Lee G, Ham B, Boyanov MI, Kemner KM, O’Loughlin EJ. 2015.

Geochemical characteristics and microbial community composition in toxic metal-rich

173

sediments contaminated with Au-Ag mine tailings. J Hazard Mater 296:147–157.

4. Sun MY, Dafforn KA, Johnston EL, Brown M V. 2013. Core sediment bacteria drive

community response to anthropogenic contamination over multiple environmental gradients.

Environ Microbiol 15:2517–2531.

5. Jurburg SD, Nunes I, Brejnrod A, Jacquiod S, Priemé A, Sørensen SJ, Van Elsas JD, Salles JF.

2017. Legacy effects on the recovery of soil bacterial communities from extreme temperature

perturbation. Front Microbiol 8:1–13.

6. Jacquiod S, Cyriaque V, Riber L, Al-soud WA, Gillan DC, Wattiez R, Sørensen SJ. 2018.

Long-term industrial metal contamination unexpectedly shaped diversity and activity response

of sediment microbiome. J Hazard Mater 344:299–307.

7. Ni C, Horton DJ, Rui J, Henson MW, Jiang Y, Huang X, Learman DR. 2016. High

concentrations of bioavailable heavy metals impact freshwater sediment microbial

communities. Ann Microbiol 66:1003–1012.

8. Griffiths BS, Philippot L. 2013. Insights into the resistance and resilience of the soil microbial

community. FEMS Microbiol Rev 37:112–129.

9. Røder HL, Hansen LH, Sørensen SJ, Burmølle M. 2013. The impact of the conjugative IncP-1

plasmid pKJK5 on multispecies biofilm formation is dependent on the plasmid host. FEMS

Microbiol Lett 344:186–192.

10. Cyriaque V, Jacquiod S, Riber L, Abu Al-soud W, Gillan DC, Sørensen SJ, Wattiez R. 2020.

Selection and propagation of IncP conjugative plasmids following long-term anthropogenic

metal pollution in river sediments. J Hazard Mater 382:121173.

11. Nishida H, Oshima T. 2019. DNA Traffic in the Environment. Singapore.

12. Perry JA, Wright GD. 2013. The antibiotic resistance “mobilome”: Searching for the link

between environment and clinic. Front Microbiol 4:1–7.

13. Pal C, Bengtsson-Palme J, Kristiansson E, Larsson DGJ. 2015. Co-occurrence of resistance

genes to antibiotics, biocides and metals reveals novel insights into their co-selection potential.

BMC Genomics 16:1–14.

14. Bahl MI, Hansen LH, Sørensen SJ. 2009. Persistence Mechanisms of Conjugative Plasmids, p.

174

. In Al., MBG et (ed.), Horizontal Gene Transfer: Genomes in Flux. Humana Press.

15. Lopatkin AJ, Meredith HR, Srimani JK, Pfeiffer C, Durrett R, You L. 2017. Persistence and

reversal of plasmid-mediated antibiotic resistance. Nat Commun 8:1689.

16. Kottara A, Hall JPJ, Harrison E, Brockhurst MA. 2018. Variable plasmid fitness effects and

mobile genetic element dynamics across Pseudomonas species. FEMS Microbiol Ecol 94:1–7.

17. Monchy S, Benotmane MA, Janssen P, Vallaeys T, Taghavi S, Van Der Lelie D, Mergeay M.

2007. Plasmids pMOL28 and pMOL30 of Cupriavidus metallidurans are specialized in the

maximal viable response to heavy metals. J Bacteriol 189:7417–7425.

18. Smets BF, Morrow JB, Pinedo CA. 2003. Plasmid Introduction in Metal-Stressed , Subsurface-

Derived Microcosms : Plasmid Fate and Community Response. Appl Environ Microbiol

69:4087–4097.

19. Parra B, Tortella GR, Cuozzo S, Martínez M. 2019. Negative effect of copper nanoparticles on

the conjugation frequency of conjugative catabolic plasmids. Ecotoxicol Environ Saf 169:662–

668.

20. Klümper U, Dechesne A, Riber L, Brandt KK, Gülay A, Sørensen SJ, Smets BF. 2017. Metal

stressors consistently modulate bacterial conjugal plasmid uptake potential in a

phylogenetically conserved manner. ISME J 11:152–165.

21. Bahl MI, Sørensen SJ, Hansen LH. 2004. Quantification of plasmid loss in Escherichia coli

cells by use of flow cytometry. FEMS Microbiol Lett 232:45–49.

22. Gougoulias C, Shaw LJ. 2012. Evaluation of the environmental specificity of Fluorescence In

Situ Hybridization (FISH) using Fluorescence-Activated Cell Sorting (FACS) of probe

(PSE1284)-positive cells extracted from rhizosphere soil. Syst Appl Microbiol 35:533–540.

23. Yamaguchi N, Ohba H, Nasu M. 2006. Simple detection of small amounts of Pseudomonas

cells in milk by using a microfluidic device. Lett Appl Microbiol 43:631–636.

24. Lawson TS, Connally RE, Vemulpad S, Piper JA. 2012. Dimethyl formamide-free, urea-NaCl

fluorescence in situ hybridization assay for Staphylococcus aureus. Lett Appl Microbiol

54:263–266.

25. Kommerein N, Stumpp SN, Musken M, Ehlert N, Winkel A, Haussler S, Behrens P, Buettner

175

FFR, Stiesch M. 2017. An oral multispecies biofilm model for high content screening

applications. PLoS One 12:1–21.

26. Oksanen J, Blanchet FG, Friendly M, Kindt R, Legendre P, McGlinn D, Minchin PR, O’Hara

RB, Simpson GL, Solymos P, Stevens . Henry H., Szoecs E, Wagner H. 2019. vegan:

Community Ecology Package. R package version 2.5-6.

27. Juggins S. 2019. rioja: Analysis of Quaternary Science Data. 0.9-22.

28. Warnes GR, Bolker B, Bonebakker L, Gentleman R, Liaw WHA, Lumley T, Maechler M,

Magnusson A, Moeller S, Schwartz M, Venables B. 2016. Package “gplots”: Various R

programming tools for plotting data. R Packag version 2170 1–68.

29. Leroy B, De Meur Q, Moulin C, Wegria G, Wattiez R. 2015. New insight into the

photoheterotrophic growth of the isocytrate lyase-lacking purple bacterium rhodospirillum

rubrum on acetate. Microbiol (United Kingdom) 161:1061–1072.

30. Taghavi S, Lesaulnier C, Monchy S, Wattiez R, Mergeay M, Lelie D. 2009. Lead(II) resistance

in Cupriavidus metallidurans CH34: Interplay between plasmid and chromosomally-located

functions. Antonie van Leeuwenhoek, Int J Gen Mol Microbiol 96:171–182.

31. Cortes L, Wopereis H, Tartiere A, Piquenot J, Gouw JW, Tims S, Knol J, Chelsky D. 2019.

Metaproteomic and 16S rRNA gene sequencing analysis of the infant fecal microbiome. Int J

Mol Sci 20:9–12.

32. Xu FF, Imlay JA. 2012. Silver(I), mercury(II), cadmium(II), and zinc(II) target exposed

enzymic iron-sulfur clusters when they toxify Escherichia coli. Appl Environ Microbiol

78:3614–3621.

33. Wang P, Zhang S, Wang C, Lu J. 2012. Effects of Pb on the oxidative stress and antioxidant

response in a Pb bioaccumulator plant Vallisneria natans. Ecotoxicol Environ Saf 78:28–34.

34. Chenier D, Beriault R, Mailloux R, Baquie M, Abramia G, Lemire J, Appanna V. 2008.

Involvement of fumarase C and NADH oxidase in metabolic adaptation of Pseudomonas

fluorescens cells evoked by aluminum and gallium toxicity. Appl Environ Microbiol 74:3977–

3984.

35. Jarosławiecka A, Piotrowska-Seget Z. 2014. Lead resistance in micro-organisms. Microbiol

176

(United Kingdom) 160:12–25.

36. Kurth C, Schieferdecker S, Athanasopoulou K, Seccareccia I, Nett M. 2016. Variochelins,

Lipopeptide Siderophores from Variovorax boronicumulans Discovered by Genome Mining. J

Nat Prod 79:865–872.

37. Naik MM, Dubey SK. 2011. Lead-enhanced siderophore production and alteration in cell

morphology in a Pb-resistant Pseudomonas aeruginosa strain 4EA. Curr Microbiol 62:409–

414.

38. Bell A, Hancock REW. 1989. Outer membrane protein H1 of Pseudomonas aeruginosa:

Purification of the protein and cloning and nucleotide sequence of the gene. J Bacteriol

171:3211–3217.

39. Pinilla-Redondo R, Cyriaque V, Jacquiod S, Sørensen SJ, Riber L. 2018. Monitoring plasmid-

mediated horizontal gene transfer in microbiomes: recent advances and future perspectives.

Plasmid 99:56–67.

40. Bahl MI, Hansen LH, Sørensen SJ. 2007. Impact of conjugal transfer on the stability of IncP-1

plasmid pKJK5 in bacterial populations. FEMS Microbiol Lett 266:250–256.

41. Hynninen A, Touzé T, Pitkänen L, Mengin-Lecreulx D, Virta M. 2009. An efflux transporter

PbrA and a phosphatase PbrB cooperate in a lead-resistance mechanism in bacteria. Mol

Microbiol 74:384–394.

42. Bates S, Cashmore AM, Wilkins BM. 1998. IncP plasmids are unusually effective in mediating

conjugation of Escherichia coli and Saccharomyces cerevisiae: Involvement of the Tra2 mating

system. J Bacteriol 180:6538–6543.

43. Schröder G, Krause S, Zechner EL, Traxler B, Yeo HJ, Lurz R, Waksman G, Lanka E. 2002.

TraG-like proteins of DNA transfer systems and of the Helicobacter pylori type IV secretion

system: Inner membrane gate for exported substrates? J Bacteriol 184:2767–2779.

44. Lanka E, Barth PT. 1981. Plasmid RP4 specifies a deoxyribonucleic acid primase involved in

its conjugal transfer and maintenance. J Bacteriol 148:769–781.

177

Supporting method Valentine Cyriaque, Jonas Stenløkke Madsen, Laurence Fievez, Baptiste Leroy, Lars Hansen, David C. Gillan, Fabrice Bureau, Søren J. Sørensen, Ruddy Wattiez «Lead drives complex dynamics of conjugative plasmid in bacterial community»

Plasmid construction and transformation DNA fragment including the operon pbrTRABCD and tetracycline resistance gene tetR (tetC) was inserted in the plasmid pKJK5-gfpmut3-kanR. The operon pbrTRABCD was amplified by

PCR using genomic DNA of Cupriavidus metallidurans CH34 and tetR from pBR322 as templates. The two PCR product were fused by Overlap-Extension PCR (OE-PCR) DNA and plasmid extractions were carried out using DNA Extraction Genomic mini kit (A&A

Biotechnology) and Plasmid Midi AX kit (A&A Biotechnology), respectively following manufacturer instructions. PCR amplification of the pbrTRABCD (7091 bb) was done using primers pbrD_xbaI (5’-CATACTTCTAGACTACCTACAGGCGTAGGCAC-3’) and pbrTOLtet (5’-GGAGAACTGTGAATGCGCAGGCGTTACACCTGGGTAGAT-3’). PCR amplification of tetR gene (1221bp) was insured by primers Tet_SacI (5’-

CGTCATGAGCTCAGGCCCTTTCGTCTTCAAGA-3’) and Tet_TOLpbr (5’-

ATCTACCCAGGTGTAACGCCTGCGCATTCACAGTTCTCC-3’). PCRs were carried with Phusion™ Hot Start II DNA Polymerase (ThermoFisher Scientific) following the manufacturer instructions (DMSO included). Melting temperature and elongation time for pbrTRABCD amplification were 61°C and 3.5 minutes. Melting temperature and elongation time for tetR amplification were 63°C and 40 seconds. pbrTRABCD PCR product was loaded on an electrophoresis gel (10 g/L agarose; 1.5 kV, 60 minutes) and corresponding agar band was cut for DNA extraction using QIAEX II Gel Extraction Kit (Qiagen). tetR PCR products was extracted using the QIAquick PCR Purification Kit (Quiagen). OE-PCR was performed using 50 ng of both DNA fragments, Tet_SacI and pbrD_xbaI primers and the Phusion™ Hot

Start II DNA Polymerase (ThermoFisher Scientific) following the manufacturer instructions

(DMSO included). Melting temperature and elongation time were 62°C and 3.5 minutes.

Obtained PCR product was loaded on an electrophoresis gel (10 g/L agarose; 1.5 kV,

60 minutes) and DNA fragment (≈8 kb) was purified from the corresponding agarose band using QIAEX II Gel Extraction Kit (Qiagen). Obtained DNA fragment and the pLENTTc -

178

TcS::MCS (see following section) plasmid were cut separately using XbaI and SacI restriction enzymes (NEB; Cutsmart buffer, 90 minutes, 37°C ; stop reaction at 20 minutes, 65°C). DNA fragment (120 ng) and linearized plasmid (40 ng) were ligated together using T4 ligase (NEB) at 16°C overnight (stop reaction 65°C, 10 minutes). The obtained plasmid was introduced in

Electrocomp™ GeneHogs® E. coli by electroporation. For that, 1 µL of DNA was gently mixed with 50 µL of electrocompetent cells and transferred into a 1 mm BioRad cuvette.

Electroporation was carried out using MicroPulser™ Electroporator (BioRad) with 1.80 kV, 1 pulse. Electroporated cells were incubated for 2 hours at 30°C, 300 RPM in SOC medium and clones were selected on tetracycline 20 µg/mL LB agar. The obtained plasmid was cut with the restriction enzyme BglII (NEB; Buffer 3, 90 minutes, 37°C; stop reaction at 20 minutes, 65°C) and the fragment of interest (Mu Ends, pbrTRABCD and tetR fragments, ≈8 kb) was loaded on an electrophoresis gel (10 g/L agarose; 1.5 kV, 60 minutes). The corresponding agar band was cut for DNA extraction using QIAEX II Gel Extraction Kit (Qiagen). The pbrTRABCD-tetR fragment was inserted into pKJK5-gfpmut3-kanR by MuA transposition using 2 µg of the plasmid and 50 ng of the BglII cut fragment. The obtained plasmid was electroporated in

Electrocomp™ GeneHogs® E. coli as previously described. Clones were selected on tetracycline 20 µg/mL LB agar and 20 clones were grown in LB supplemented with either tetracycline (20 µg/mL) or tetracycline (20 µg/mL) and trimethoprim (50 µg/mL). One clone growing in the presence of tetracycline but not in the presence of trimethoprim was selected.

Construction was checked by PCR from the purified plasmid. The strain Pseudomonas putida

R R R KT2440::PlppmCherry-kan //pKJK5-gfpmut3-pbrTRABCD-kan -tet was obtained by

R conjugation assay between Pseudomonas putida KT2440::PlppmCherry-kan and GeneHogs®

E. coli//pKJK5-gfpmut3-pbrTRABCD-kanR-tetR. For that, 2 mL of overnight pre-cultures of both strains were washed twice in 2 mL of hot LB (2 minutes, 7000×g, 37°C) and 150 µL of each were mixed in 600 µL of hot LB (37°C). The mixture was washed twice in 1 mL and

179

resuspended in 30 µL of hot LB that was deposited on pre-heated LB-agar plate and incubated

5 hours at 37°C. The dried cell mixed was scrapped off and then resuspended in 200 µL of NaCl solution (9 g/L). Then,100 µL were diluted in 5 mL LB supplemented with tetracycline (50

µg/mL). Grown clones were isolated on LB-agar supplemented with tetracycline (50 µg/mL). pLENTTc -TcS::MCS construction pLENTTc1 (Fig 1) was constructed from the pACYC177 backbone. pACYC177 was digested with StuI and a 2600 bp band comprising the p15A origin of replication and the bla gene conferring ampicillin resistance was excised from an agarose gel and purified using QiexII

(Qiagen). This was ligated to a 1599 bp PCR fragment that was generated with the primers

EntklonFw 5’-GAC GTT GTA AAA CGA CGG CCA G-3’ and EntklonRev 5’-GAA ACA

GCT ATG ACC ATG ATT ACG CC-3’, with pEntranceposon (tet) as template. This fragment contains the tetracycline resistance entranceposon cassette with the tetC gene and surrounding BglII sites flanking the MuA ends. The ligation mix was electroporated into E. coli Genehogs and transformants were selected on LB agar with ampicillin (100µg/ml) and tetracycline (10µg/ml). The pLENTTc1 construct was then verified using BglII restriction enzyme digests followed by electrophoresis and by sanger sequencing out from the entranceposon part with two different primers EnttetFW (5’- GTC AAA CAT GAG AAG GAT

CCG-3’) and SeqA (5’- ATC AGC GGC CGC GAT C-3’). From the sequencing reaction the orientation of the insert could be established (Figure 1). pLENTTc and pUCP22not plasmids were cut separately using HindIII and NotI HF (pLENTTc) or NotI HF (pUCP22not) restriction enzymes (NEB; Cutsmart buffer, 90 minutes, 37°C ; stop reaction at 20 minutes, 65°C). pLENTTc1 abd MCS obtained fragments were ligated using T4 ligase (NEB) at 16°C overnight

(stop reaction 65°C, 10 minutes). The obtained plasmid was introduced in Electrocomp™

GeneHogs® E. coli by electroporation as described in the main text.

180

Figure 1 : pLENTTc map

V. paradoxus B4 and D. acidovorans SPH-1 transformation with pKJK5 plasmids for burden assay

From overnight pre-cultures of each plasmid recipient and donor cell cultures (GeneHogs® E. coli- pKJK5-gfp or GeneHogs® E. coli- pKJK5-gfp-pbr), 2 mL were washed twice in 2 mL of hot LB (2 minutes, 7000×g, 37°C) and 150 µL were mixed in 600 µL of hot LB (37°C). The mix was washed twice in 1 mL and resuspended in 30 µL of hot LB that was dropped on pre- heated LB-agar plate incubated overnight at 37°C. The dried cell mixed was scrapped, resuspended in 200 µL of NaCl (9g/L) whose 100 µL were spread and incubated overnight either on LB plate supplemented with ampicillin (100 µg/mL) and tetracycline (20 µg/mL) for

D. acidovorans selection or on 457- agar medium supplemented with tetracycline (20 µg/mL) for V. paradoxus selection to insure bacteria carry the plasmid.

SWATH metaproteomic analysis

Tryptic peptides were separated on a C18 column (Acclaim PepMap100, 3 μm, 150 μm × 25 cm, Dionex) with a linear acetonitrile gradient (5 to 35% of acetonitrile (v/v), 0,1% FA, 300

181

nl.min-1, 120 min) in water containing 0.1% (v/v) formic acid. MS survey scans (m/z 400-

1250, 100 ms accumulation time) were succeeded by 50 SWATH acquisition overlapping windows covering the precursor m/z range. Ion collision induced dissociation were carried on using rolling collision energy, and fragment ion were accumulated for 95 ms in high sensitivity mode. SWATH technology identifies obtained spectra by comparing them to a referential spectral library built by Data-Dependent Acquisition (DDA). To build the library, proteins were extracted and digested as previously described, from monoculture of each used strain and cocultures (Table S1) at lead concentration of 0 mM and 1 mM. Parameters used to acquire the

DDA spectra were as follow: MS scan (m/z 400-1500, 500 ms accumulation time) followed by

50 MS/MS scans (m/z 100-1800, 50 ms accumulation time, intensity threshold at 200 c.p.s).

AB Sciex ProteinPilot™ 4.5 software was used to process the DDA mass spectrometry data.

Spectra identification was performed by searching against the corresponding strain UniProt entries with parameters including carbamidomethyl cysteine, oxidized methionine, all biological modifications, amino acid substitutions and missed cleavage site. The final SWATH reference spectral library referred to the proteins identified at a false discovery rate below 1%.

182

Supporting Figure File Valentine Cyriaque, Jonas Stenløkke Madsen, Laurence Fievez, Baptiste Leroy, Lars Hansen, David C. Gillan, Fabrice Bureau, Søren J. Sørensen, Ruddy Wattiez «Lead drives complex dynamics of conjugative plasmid in bacterial community»

Figure S1: Example of results from flow cytometry displaying V. paradoxus B4, D. acidovorans SPH-1, P. putida KT2440 strains in pure-cultures and a two-members community gathering V. paradoxus B4 and P. putida KT2440. FACS (forward scatter 390 V, side scatter 176 V, detectors for green fluorescence associated to GFPmut fluorescence (FITC, bandpass filter 530/30 nm, 501 V) and for red fluorescence associated to AlexaFluor 647 (APC-A, bandpass filter 670/14 nm, 550 V).

183

Figure S2: Patterns of differential V. paradoxus B4’s proteins after normalisation on summed area of all proteins for each sample (A), and cell ratio in the two-members community obtained by flow cytometry (Cell normalized, B) or complemented with an normalization on summed area of all proteins belonging to proteins attributed to the specific strain expressions (Normalized of total proteins identified as specific strain, C). Volcano plot display the log2 of the fold-change ratios and the negative log10 of the p-value. Dot lines demarcates thresholds of fold changes below 0,66 (log2(0,66) =-0,58) and over 1,5 (log2(0,66) =0,58) and p-value below 0,05 (-log10(0,05) =1,30).

184

Figure S3: Patterns of differential P. putida KT2440’s proteins (when associated to V. paradoxus B4) after normalisation on summed area of all proteins for each sample (A), and cell ratio in the two-members community obtained by flow cytometry (Cell normalized, B) or complemented with an normalization on summed area of all proteins belonging to proteins attributed to the specific strain expressions (Normalized of total proteins identified as specific strain, C). Volcano plot display the log2 of the fold-change ratios and the negative log10 of the p-value. Dot lines demarcates thresholds of fold changes below 0,66 (log2(0,66) =-0,58) and over 1,5 (log2(0,66) =0,58) and p-value below 0,05 (-log10(0,05) =1,30).

185

Figure S4: Patterns of differential D. acidovorans SPH-1’s proteins after normalisation on summed area of all proteins for each sample (A), and cell ratio in the two-members community obtained by flow cytometry (Cell normalized, B) or complemented with an normalization on summed area of all proteins belonging to proteins attributed to the specific strain expressions (Normalized of total proteins identified as specific strain, C). Volcano plot display the log2 of the fold-change ratios and the negative log10 of the p-value. Dot lines demarcates thresholds of fold changes below 0,66 (log2(0,66) =-0,58) and over 1,5 (log2(0,66) =0,58) and p-value below 0,05 (-log10(0,05) =1,30).

186

Figure S5: Patterns of differential P. putida KT2440’s proteins (when associated to D.acidovorans SPH- 1) after normalisation on summed area of all proteins for each sample (A), and cell ratio in the two- members community obtained by flow cytometry (Cell normalized, B) or complemented with an normalization on summed area of all proteins belonging to proteins attributed to the specific strain expressions (Normalized of total proteins identified as specific strain, C). Volcano plot display the log2 of the fold-change ratios and the negative log10 of the p-value. Dot lines demarcates thresholds of fold changes below 0,66 (log2(0,66) =-0,58) and over 1,5 (log2(0,66) =0,58) and p-value below 0,05 (- log10(0,05) =1,30).

187

Figure S6: MRM quantification of PbrA protein in two-members communities (using either Variovorax paradoxus B4 (A) or D. acidovorans SPH-1 (B) as plasmid recipient) after normalisation by the proportion of plasmid carrying cells (donors and transconjugants) obtained by flow cytometry. Significance was obtained with ANOVA followed by Tuckey test. . p-value<0,07; *: p-value<0,05; **: p-value<0,01; ***: p-value<0,001 (n=3).

Figure S7: Supplemental heatmaps are displayed in supplemental annexed specific document. Heatmaps were built with centred-scaled log-2 transformed abundances of proteins classified in COGs, using Euclidean distance and average clustering.

188

Figure S8: Abundance of conjugation involved proteins (>1 identified peptide) encoded by the pKJK5 plasmid

(normalisation on summed area of all peptides for each sample) displayed on a heatmap (Euclidean distance and average clustering) as extracted from two-members communities using either Variovorax paradoxus B4 (A) or D. acidovorans SPH-1 (B) as plasmid recipient. Legend: Green bars concern two-members community without the pbrTRABCD operon (G). Blue bars concern two-members community including the pbrTRABCD operon (GP).

The number indicates the Pb (II) concentration (mM). p-values were calculated from log-2 transformed abundances using t-test.

189

Heatmap Supporting Information Valentine Cyriaque, Jonas Stenløkke Madsen, Laurence Fievez, Baptiste Leroy, Lars Hansen, David C. Gillan, Fabrice Bureau, Søren J. Sørensen, Ruddy Wattiez «Lead drives complex dynamics of conjugative plasmid in bacterial community» Functional responses of P. putida KT2440, V. paradoxus and D. acidovorans SPH-1 Description and heatmaps

Variovorax paradoxus B4 as recipient strain When the plasmid donor P. putida KT2440 was grown in the presence of V. paradoxus

B4, a total of 632 proteins assigned to P. putida KT2440 were significantly impacted by lead

(p-value<0.05; 233, p-value<0.01). A total of 734 proteins assigned to V. paradoxus B4 were significantly impacted by lead (p-value<0.05; 296, p-value<0,01).

When the plasmid donor P. putida KT2440 was grown in the presence of V. paradoxus

B4, proteins using Fe (II) (e.g. Cytochrome b) or. [4 Fe-4S] (e.g. putative oxidoreductase or succinate dehydrogenase iron-sulfur subunit (Figure 4), Fumarate hydratase class I, C COG) as co-factor were underabundant and in some cases, replaced by an alternative class of proteins

(e.g. Fumarate hydratase class II, Figure S8). Proteins using Mg(II) (e.g. isocitrate lyase, isocitrate dehydrogenase [NADP], C COG; Adenylosuccinate synthetase, formate-dependent phosphoribosylglycinamide formyltransferase ,F COG) were also underabundant. Guanine deaminase (F COG), binding 1 zinc ion per subunit was underabundant at 1,5 mM of lead. Most translation, ribosomal structure and modification associated proteins (J COG) were negatively impacted by 1,5 mM of lead. Abundance of ClpV1 chaperone, peroxidases and endopeptidase

(O COG) were increased with lead as well as helicases and exonucleases proteins (L COG).

The membranes, especially from bacteria carrying the pbrTRABCD operon, were enriched in a penicillin binding protein 1B as well as toluene efflux pumps (Figure 4) and nucleotide sugar associated proteins (murG associated protein) in pbrTRABCD free cells (M COG). Membranes were also enriched in a pyoverdine ABC export system (V COG) and an efflux pump membrane transporter (P COG). Phosphate binding and transport proteins as well as efflux transporters (P

COG) were superabundant at 1,5 mM of lead. The abundance of iron related proteins (e.g. heme oxygenase, outer membrane ferripyoverdine and heme receptors, P COG) were increased from

1 mM of lead concentration.

190

Impacts in V. paradoxus B4 protein abundances were as follow. Abundance of proteins using Zn (II) (e.g. Putative zinc-type alcohol dehydrogenase, C COG; ATP dependent zinc metalloprotease FtsH, O COG), Fe(II) (e.g. Cbb3-type cytochrome c oxidase subunit, C COG) or [4 Fe-4S] cluster (e.g. ferredoxin, respiratory nitrate reductase, succinate dehydrogenase,

Putative formate dehydrogenase FdhB, fumarate hydratase class I, C COG) as co-factor was decreased while the abundance of the fumarate hydratase class II was increased from a concentration of 0.5 mM of lead (Figure 4). Abundance of isocitrate dehydrogenase using

Mg(II) as cofactor was decreased unlike pyruvate dehydrogenase also using Mg(II) as cofactor.

All translation, ribosomal structure and modification associated proteins (J COG) and most nucleotide and amino acid transport and metabolism associated proteins (E and F COGs) were negatively impacted by 1,5 mM of lead. Abundance of proteins involved in carbohydrate import was increased (e.g. putative sugar ABC transporters, G COG). A putative MotA/TolQ/ExbB proton channel protein (U COG), Type VI secretion system (O COG) and pilus assembly proteins (type IV system and Pil proteins, N COGs) abundances were significantly increased by 1,5 mM of lead. The glutathione-S-transferase abundance was increased at 1 mM of lead.

Membrane was also strengthened by Omp proteins (M COG). Phosphate binding and transport involved proteins, catalase peroxidases, RND family efflux transporters and putative TRAP decarboxylase transporter subunits (P COG) displayed an increased abundance at 1.5 mM of lead while the abundance of TonB dependent siderophores and a ferrous iron transport protein

B (P COG) were increased from a lead concentration of 1 mM.

191

Delftia acidovorans SPH-1 as recipient strain When the P. putida KT2440 plasmid donor was grown in the presence of D. acidovorans

SPH-1, a total of 739 proteins were significantly impacted by metals (p-value<0.05; 352, p- value<0.01). Abundance of proteins identified as D. acidovorans SPH-1 cells, significantly impacted by metals (585, p-value<0.05; 254, p-value<0,01) were mostly increased at a lead concentration of 1.5 mM (Figure S10)

When the P. putida KT2440 plasmid donor was grow in the presence of D. acidovorans

SPH-1, proteins using [4 Fe-4S] (e.g. Fumarate hydratase class I, C COG; Aconitate hydratase

B, C COG; or succinate dehydrogenase (Figure 4), C COG) as co-factor were underabundant and in some cases, replaced by an alternative classis of proteins (e.g. Fumarate hydratase class

II, Figure S8). Proteins using Mg(II) (e.g. ATP synthase , malic enzyme B or isocitrate dehydrogenase , C COG; adenosyl succinate synthase, F COG) were also underabundant.

Proteins using Zn (II) (e.g. aldolases, G COG; alcohol dehydrogenase, C COG; 30s ribosomal protein S2, J COG), Mn(II) (Poly-A polymerase I) or Co(II) (e.g. corrinoid adenosyl transferase,

H COG) co-factors were superabundant. Lipid stock proteins (PHA synthase 2, Long-chain- fatty-acid/CoA ligase, I COG), endopeptidases, (O COG), membrane associated protein chaperone SurA (O COG), and AlgQ transcriptional regulatory protein were superabundant at

1,5 mM of lead. Membranes were enforced with superabundant lipoproteins, and nucleotide sugar associated proteins, OmpA family proteins, Penicillin binding protein1B (Figure 4) and outer-membrane efflux proteins (M COG) while porins were underabundant (M COG).

Membranes were also strengthened by superabundant efflux proteins such as Cadmium translocating P-type ATPases CadII, Mg transporter, and efflux pump membrane trans (P

COG). The abundance of proteins involved in phosphate entry regulation PhoB (T COG) as well as in phosphate binding and transport (P COG) was increased. Finally, proteins involved against ROS damaged (e.g. Glutathione S-transferase (O COG) or Flavohemoproteins (C COG)

192

and were also superabundant at 1,5 mM while outer membrane ferripyoverdine (P COG) was over abundant at 0,5 and 1 mM.

Impacts in D. acidovorans SPH-1 protein abundances were as follow. Abundance of a few proteins using Fe(II) (e.g. Gluconate 2-dehydrogenase, C COG) or a [4 Fe-4S] cluster (e.g.

Cytochrome c oxidase accessory protein CcoG, a 4Fe-4S ferredoxin iron-sulfur binding domain protein or a molydopterin dinucleotide-binding region including protein, C COG) as co-factor were decreased with the presence of metals. However, some of them were overabundant from a lead concentration of 0,5 mM such as nitrate reductase sub-units and an oxidoreductase FAD- binding protein (C COG). This nitrate-reductase was increased along with a nitroreductase (C

COG) and D-amino-dehydrogenases (E COG). An oxidoreductase FAD/NAD(P)-binding domain protein, proteins using a [2 Fe-2S] and a FAD linked oxidase domain protein using a

[4 Fe-4S] cluster as cofactor (C COGs) were superabundant at 1.5 mM of lead as well as Mg(II) cofactor dependent proteins such as ATP synthase subunits and a succinate--CoA ligase [ADP- forming] subunit beta. The membrane was strengthened by OmpA proteins and a LPS assembly protein LptD (M COG). The abundance of the chaperone HscA (O COG) was increased at 1,5 mM of lead as well as phosphatases, phosphate binding and transport associated proteins, efflux transporter RND family proteins, heavy metal efflux pumps molybdenum ABC transporters and magnesium P-type ATPases (P COG). Other efflux transporter RND family proteins, heavy metal efflux pumps as well as TonB dependent siderophores and iron permeases were overabundant from 0,5 mM of lead (P COG). Besides, a Gamma-glutamyl transferases was increased in abundance at a lead concentration of 1,5 mM (E COG). Finally, a DNA mismatch repair protein abundance was increased when lead concentration reached 1,5 mM (L COG).

Notably, the TetR transcriptional regulator was increased with high lead concentration (K

COG).

193

Following heatmaps are gathered as Figure S10: Heatmaps were built with centred-scaled log-2 transformed abundances of proteins classified in COGs, using Euclidean distance and average clustering. COG clusters are defined as follow:

Cellular processes and signalling [D] Cell cycle control, cell division, chromosome partitioning [M] Cell wall/membrane/envelope biogenesis [N] Cell motility [O] Post-translational modification, protein turnover, and chaperones [T] Signal transduction mechanisms [U] Intracellular trafficking, secretion, and vesicular transport [V] Defence mechanisms Information storage and processing [J] Translation, ribosomal structure and biogenesis [K] Transcription [L] Replication, recombination and repair Metabolism [C] Energy production and conversion [E] Amino acid transport and metabolism [F] Nucleotide transport and metabolism [G] Carbohydrate transport and metabolism [H] Coenzyme transport and metabolism [I] Lipid transport and metabolism [P] Inorganic ion transport and metabolism [Q] Secondary metabolites biosynthesis, transport, and catabolism Poorly characterized [S] Function unknown

194